WO2025240289A1

WO2025240289A1 - Document to dialogue audio conversion with text format transformation

Info

Publication number: WO2025240289A1
Application number: PCT/US2025/028835
Authority: WO
Inventors: Krishna Asur Bharat; Keyvan AMIRI; Zhi Li
Original assignee: Google LLC
Current assignee: Google LLC
Priority date: 2024-05-13
Filing date: 2025-05-12
Publication date: 2025-11-20
Anticipated expiration: 2026-11-13

Abstract

Systems and methods for academic paper-to-dialogue audio conversion can include obtaining a document file, generating a dialogue script, and converting the dialogue script into multi-speaker speech data. The systems and methods can leverage generative language models, evaluation models, and text-to-speech models for performing the conversion. The systems and methods can include script evaluation and revision to refine the script output before audio conversion.

Description

DOCUMENT TO DIALOGUE AUDIO CONVERSION WITH TEXT FORMAT

TRANSFORMATION

PRIORITY CLAIM

[0001] The present application is based on and claims priority to United States Provisional Application 63/646,152 having a filing date of May 13, 2024. Application claims priority to and the benefit of each of such applications and incorporates all such applications herein by reference in their entirety.

FIELD

[0002] The present disclosure relates generally to document-to-speech conversion. More particularly, the present disclosure relates to generating model -generated scripts based on an academic paper and generating speech data from the model -generated scripts.

BACKGROUND

[0003] Research papers and other long form documents can be difficult to understand in their original format. Additionally, users may learn in different ways with some users having greater difficulty in learning based on reading long form documents. The structure, terminology, and medium of academic papers may not be intuitive to users, which may deter users, may cause confusion, and/or may lead to misunderstanding of the content of the paper. Moreover, some readers may struggle with long form text and may prefer other forms of media.

SUMMARY

[0004] Aspects and advantages of embodiments of the present disclosure will be set forth in part in the following description, or can be learned from the description, or can be learned through practice of the embodiments.

[0005] One example aspect of the present disclosure is directed to a computing system for document-to-dialogue audio conversion. The system can include one or more processors and one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations. The operations can include obtaining a document. The operations can include processing the document with a generative language model to generate a model- generated script. The model-generated script can include a plurality of interleaved dialogue turns between two or more speakers. The operations can include processing the document and the model -generated script with the generative language model to generate a revised script. The operations can include processing the revised script with a text-to-speech model to generate audio data. The audio data can be descriptive of the revised script being output in audio form by a first speaker voice and a second speaker voice. The operations can include providing the audio data for playback.

[0006] In some implementations, the document can include a research paper. The operations can include evaluating, with a machine-learned evaluation model, the modelgenerated script to generate an evaluation model output. In some implementations, the revised script can be generated based in part on the evaluation model output. The revised script can include a first set of dialogue associated with a first speaker and a second set of dialogue associated with a second speaker. In some implementations, the operations can include processing the revised script and the audio data to determine the second speaker voice speaks a dialogue line associated with the first speaker, generating an additional text-to- speech prompt in response to determining the second speaker voice speaks the dialogue line associated with the first speaker, and processing the revised script and the additional text-to- speech prompt with the text-to-speech model to generate second audio data. The first set of dialogue can be associated with a host role. The second set of dialogue can be associated with a technical expert role. In some implementations, the first set of dialogue can include questions. The second set of dialogue can include responses to the questions. The responses can be generated based on content of the academic paper.

[0007] In some implementations, processing the document with the generative language model to generate the model-generated script can include generating semantic understanding data based on content and structure of the academic paper, generating a first set of questions, generating a first set of responses responsive to the first set of questions based on the semantic understanding data, and interweaving the first set of questions and the first set of responses to generate a chain of dialogue. The first set of questions can be generated based on determining the academic paper is associated with a particular information type based on the semantic understanding data. In some implementations, the revised script can include a second plurality of interleaved dialogue turns between the two or more speakers. The revised script can include a second model -generated script that differs from the model-generated script based on the generative language model being conditioned on the evaluation model output. [0008] Another example aspect of the present disclosure is directed to a computer- implemented method for document-to-dialogue audio conversion. The method can include obtaining, by a computing system including one or more processors, a document. The document can include an academic paper. The method can include processing, by the computing system, the document with a generative language model to generate a modelgenerated script. The model-generated script can include a plurality of interleaved dialogue turns between two or more speakers. The method can include evaluating, by the computing system and with a machine-learned evaluation model, the model-generated script to generate an evaluation model output. The method can include processing, by the computing system, the document, the model-generated script, and the evaluation model output with the generative language model to generate a revised script. The method can include processing, by the computing system, the revised script with a text-to-speech model to generate audio data. The audio data can be descriptive of the revised script being output in audio form by a first speaker voice and a second speaker voice.

[0009] In some implementations, the method can include determining, by the computing system, the academic paper is older than a threshold age, obtaining an agedisclaimer prompt in response to determining the academic paper is older than the threshold age, and processing, by the computing system, the age-disclaimer prompt, the revised script, the model -generated evaluation output, and the document with the generative language model to generate an augmented script. The augmented script can include an augmented version of the revised script. The revised script can be augmented to include one or more disclaimers on a year of publication of the research paper.

[0010] In some implementations, evaluating, with the machine-learned evaluation model, the model -generated script to generate the evaluation model output can include processing the model -generated script and the document with the machine-learned evaluation model to determine the model-generated script includes one or more factual inaccuracies and generating, with the machine-learned evaluation model, the evaluation model output that includes one or more facts from the document. In some implementations, evaluating, with the machine-learned evaluation model, the model-generated script to generate the evaluation model output can include processing the model-generated script and the document with the machine-learned evaluation model to determine the model-generated script includes a first person perspective when referring to the document and generating, with the machine-learned evaluation model, the evaluation model output that includes instructions to include dialogue from a perspective of a reader. The generative language model can include an autoregressive language model. In some implementations, the machine-learned evaluation model can include a natural language processing model that includes a plurality of classifiers and a prompt generation model.

[0011] Another example aspect of the present disclosure is directed to one or more non-transitory computer-readable media that collectively store instructions that, when executed by one or more computing devices, cause the one or more computing devices to perform operations. The operations can include obtaining a document. The document can include an academic paper. The operations can include processing the document with a generative language model to generate a model -generated script. The model -generated script can include a plurality of interleaved dialogue turns between two or more speakers. The operations can include processing the model-generated script with a machine-learned evaluation model to generate an evaluation model output. The operations can include generating a revision prompt based on the evaluation model output. The operations can include processing the document, the revision prompt, and the evaluation model output with the generative language model to generate a revised script. The operations can include processing the revised script with a text-to-speech model to generate audio data. The audio data can be descriptive of the revised script being output in audio form by a first speaker voice and a second speaker voice.

[0012] In some implementations, the text-to-speech model can include a text-to- speech autoencoder. The text-to-speech model can include a transformer model. The operations can include obtaining a second document. The second document can include a second academic paper. The operations can include concatenating the document and the second document to generate a combined document and processing the combined document with the generative language model to generate a multi -document script.

[0013] Other aspects of the present disclosure are directed to various systems, apparatuses, non-transitory computer-readable media, user interfaces, and electronic devices. [0014] These and other features, aspects, and advantages of various embodiments of the present disclosure will become better understood with reference to the following description and appended claims. The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate example embodiments of the present disclosure and, together with the description, serve to explain the related principles. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Detailed discussion of embodiments directed to one of ordinary skill in the art is set forth in the specification, which makes reference to the appended figures, in which: [0016] Figure 1 depicts a block diagram of an example audio data generation system according to example embodiments of the present disclosure.

[0017] Figure 2 depicts a block diagram of an example paper-to-podcast system according to example embodiments of the present disclosure.

[0018] Figure 3 depicts a flow chart diagram of an example method to perform document-to-audio conversion according to example embodiments of the present disclosure. [0019] Figure 4 depicts an illustration of an example upload interface according to example embodiments of the present disclosure.

[0020] Figure 5 depicts an illustration of an example selection interface according to example embodiments of the present disclosure.

[0021] Figure 6 A depicts an illustration of an example library interface according to example embodiments of the present disclosure.

[0022] Figure 6B depicts an illustration of an example playback interface according to example embodiments of the present disclosure.

[0023] Figure 7 depicts a flow chart diagram of an example method to perform multidraft script generation according to example embodiments of the present disclosure.

[0024] Figure 8 depicts a flow chart diagram of an example method to perform revised script generation according to example embodiments of the present disclosure. [0025] Figure 9 depicts a flow chart diagram illustrating an example method for training a machine-learned model according to example implementations of aspects of the present disclosure.

[0026] Figure 10 depicts a block diagram of an example processing flow for using machine-learned model(s) to process input(s) to generate output(s) according to example implementations of aspects of the present disclosure.

[0027] Figure 11 depicts a block diagram of an example sequence processing model according to example implementations of aspects of the present disclosure.

[0028] Figure 12 depicts a block diagram of an example technique for populating an example input sequence for processing by a sequence processing model according to example implementations of aspects of the present disclosure.

[0029] Figure 13 depicts a block diagram of an example model development platform according to example implementations of aspects of the present disclosure. [0030] Figure 14 depicts a block diagram of an example training workflow for training a machine-learned model according to example implementations of aspects of the present disclosure.

[0031] Figure 15 depicts a block diagram of an inference system for operating one or more machine-learned model(s) to perform inference according to example implementations of aspects of the present disclosure.

[0032] Figure 16 depicts a block diagram of an example networked computing system according to example implementations of aspects of the present disclosure.

[0033] Figure 17 depicts a block diagram of an example computing device according to example implementations of aspects of the present disclosure.

[0034] Figure 18 depicts a block diagram of an example computing device according to example implementations of aspects of the present disclosure.

[0035] Figure 19A depicts a block diagram of an example computing system that performs document-to-dialogue audio conversion according to example embodiments of the present disclosure.

[0036] Figure 19B depicts a block diagram of an example computing system that performs document-to-dialogue audio conversion according to example embodiments of the present disclosure.

[0037] Reference numerals that are repeated across plural figures are intended to identify the same features in various implementations.

DETAILED DESCRIPTION

[0038] Generally, the present disclosure is directed to systems and methods for converting a document into a script descriptive of dialogue explaining the document and then converting the script into audio data. The systems and methods disclosed herein can leverage one or more generative language models, one or more machine-learned evaluation models, and/or one or more text-to-speech models to generate a model -generated podcast based on the content of an academic paper. In particular, the systems and methods can utilize a generative language model to generate a script that includes a model-generated interleaved chain of dialogue between two or more speakers discussing the document. The model -generated script can then be processed with a machine-learned evaluation model to determine if the modelgenerated script includes any factual errors, any perspective errors (e.g., dialogue insinuating one of the speakers is the author and/or contributed to the document), any formatting errors, and/or other possible errors. The model -generated script, the evaluation model output, and/or the document may then be processed by the generative language model to generate a revised script. The model-generated script and/or the revised script may be formatted as a script for a podcast. The revised script may then be converted into audio form via processing the revised script with a text-to-speech model. The generated audio may then be provided for playback. [0039] An academic paper to podcast dialogue conversion feature may be provided as a dedicated application and/or web platform for receiving input documents and outputting audio files with the generated podcast. Alternatively and/or additionally, the conversion feature may be provided by publishing websites and/or search engines (e.g., scholastic search engines) to provide alternative mediums for consuming the content. Users may leverage the conversion feature for staying up-to-date on scientific advancements, for learning more on a topic, for understanding complex documents, and/or for other uses.

[0040] Academic papers and other long form documents can be difficult to understand in their original format. Additionally, users may learn in different ways with some users having greater difficulty in learning based on reading long form documents. The structure, terminology, and medium of academic papers may not be intuitive to users, which may deter users, may cause confusion, and/or may lead to misunderstanding of the content of the paper. Moreover, some readers may struggle with long form text and may prefer other forms of media.

[0041] A document (e.g., an academic paper or other long form documents) can be processed with a generative model (e.g., a large language model (LLM)) to perform semantic understanding of the document and generate a script of dialogue that discusses the document based on the semantic understanding. The script may then be processed with another machine-learned model to identify potential errors with factuality, perspective, and/or other potential errors. The script, document, and second model output can then be processed with the generative model (e.g., the LLM) to generate a revised script. The revised script can then be processed with a text-to-speech model to generate a dialogue audio file for output. The document to script conversion may be performed based on a tailored prompt being processed with the document by the generative model (e.g., the LLM).

[0042] Podcasts are a popular medium for some users to consume information, which includes information that may be educational in nature (e.g., a podcast on history and/or science). The paper-to-podcast conversion can provide a different medium for users to obtain information that may be more understandable and/or accessible than the original paper format. Additionally, users can now consume the information on the go, which may include listening during their commute, during a workout, and/or other activities. [0043] Although the present disclosure discusses the paper-to-podcast conversion, the systems and methods disclosed herein can obtain and process a variety of different input data types or sizes to generate a transcript and/or audio data descriptive of dialogue on a topic. For example, the systems and methods disclosed herein may determine a particular topic is trending, may obtain one or more content items (e.g., a few articles on the topic, an academic paper on the topic, a set of social media posts, a multimedia web page on the topic, etc.), may process the one or more content items to generate a transcript descriptive of a dialogue on the trending topic, and may then render an audio podcast based on the transcript. Additionally and/or alternatively, the systems and methods may obtain a plurality of content item inputs by the user and may then render a transcript and audio file based on understanding the content of the plurality of content items and based on a determined connectivity between the plurality of content items.

[0044] In some implementations, a user may input a search query. The systems and methods disclosed herein may determine a plurality of content items responsive to the search query. The search query and at least a subset of the plurality of content items can then be processed to generate a transcript. The transcript can then be processed to generate the audio file.

[0045] In some implementations, user data may be processed to determine topics of interest, level of expertise in one or more fields, tone, dialect, lexicon, and/or other features for personalizing the dialogue. For example, if a user is an expert in the field of astronomy, the systems and methods may generate more academic and concise podcasts when an astronomy paper is the input. Additionally and/or alternatively, a user’s search history, browsing history, post history, and/or purchasing history may be utilized to identify content items to be processed for generating a personalized proactive podcast output.

[0046] The system may allow for personalization based on providing a user interface that allows a user to select the tone, length, pace, voices (e.g., utilizing speaker transfer to render the text to audio), language, and/or other features.

[0047] The systems and methods of the present disclosure provide a number of technical effects and benefits. As one example, the system and methods can provide a paper- to-podcast conversion. In particular, the systems and methods can obtain a document file and generate an audio file descriptive of a model -generated podcast discussing the document. The synthetic podcast can include multiple speakers with one speaker acting as a host (or interviewer) with one or more other speakers being the technical expert that responds to the questions posed by the host. The text-to-speech model can render the dialogue lines of the model-generated script with different voices depending on which speaker the dialogue line is assigned to within the script. The paper-to-podcast conversion can leverage a generative language model, an evaluation model, and/or a text-to-speech model.

[0048] Another technical benefit of the systems and methods of the present disclosure is the ability to leverage one or more machine-learned models to understand text within an academic paper, generate a script, then generate audio with distinctive voices. The process can include performing document understanding, performing script generation based on the understanding, and then transforming the script into a different form of media (i.e., audio). [0049] Another example of technical effect and benefit relates to improved computational efficiency and improvements in the functioning of a computing system. For example, the systems and methods disclosed herein can leverage parallel processing and/or one or more lightweight models for reduced latency and for reducing computational cost for the conversion. The parallel processing can reduce latency by generating different portions of the script simultaneously. The lightweight models can provide more compact and less computationally expensive processing.

[0050] With reference now to the Figures, example embodiments of the present disclosure will be discussed in further detail.

[0051] Figure 1 depicts a block diagram of an example audio data generation system 100 according to example embodiments of the present disclosure. In some implementations, the audio data generation system 100 is configured to receive, and/or obtain, a set of input data descriptive of a document 102 and, as a result of receipt of the input data descriptive of the document 102, generate, determine, and/or provide output data that includes audio data 110 that is descriptive of two or more speakers discussing the content of the document 102. Thus, in some implementations, the audio data generation system 100 can include a generative model 104 that is operable to generate model-generated script 106 based on the content of the document.

[0052] In particular, the audio data generation system 100 can obtain a document 102. The document 102 can include a research paper, an academic paper, experimental results, a textbook, a novel, a financial report, an article, a blog post, and/or other document. The document 102 may be obtained in a portable document format. The document 102 may include structured text, tables, images, and/or embedded data.

[0053] The generative model 104 (e.g., a generative language model (e.g., a large language model)) can process the document 102 to generate a model -generated script 106. The generative model 104 may include a natural language processing model that was trained on a plurality of training datasets for a plurality of different downstream tasks. One or more of the training datasets may include long form podcast data. The model -generated script 106 may have been generated by generating a semantic understanding of the document 102, then generating a multi -turn dialogue discussing the content of the document 102 based on the semantic understanding. The model -generated script 106 can be descriptive of a podcast script that includes a host that poses questions and a technical expert that answers the questions. The answers to the questions can be generated based on the semantic understanding of the document 102.

[0054] A text-to-speech model 108 can then process the model-generated script 106 to generate audio data 110. The audio data 110 can be descriptive of the model -generated script 106 being spoken by two or more different speaker voices. The audio data 110 may be configured to be played with one or more audio player interfaces. The text-to-speech model 108 may include one or more variational autoencoders and/or one or more transformer models.

[0055] Figure 2 depicts a block diagram of an example paper-to-podcast system 200 according to example embodiments of the present disclosure. The paper-to-podcast system 200 is similar to audio data generation system 100 of Figure 1 except that paper-to-podcast system 200 further includes an evaluation model 212 to evaluate the model-generated script 206, which can be utilized to perform guided revisions of the model-generated script.

[0056] In particular, the paper-to-podcast system 200 can obtain a document 202. The document 202 can include an academic paper, a research paper, experimental results, a textbook, a novel, a financial report, an article, a blog post, and/or other document. The document 202 may be obtained in a portable document format, an application-specific document file, a hypertext markup language file, and/or other file format. The document 202 may include structured text, tables, images, code, and/or embedded data.

[0057] The generative model 204 (e.g., a generative language model (e.g., a vision language model)) can process the document 202 to generate a model -generated script 206. The generative model 204 may include a natural language processing model that was trained on a plurality of training datasets (e.g., articles, novels, conversations, etc.) for a plurality of different downstream tasks (e.g., filling in the blank, sequence prediction, article generation, image captioning, etc.). One or more of the training datasets may include long form podcast data (e.g., transcripts from a plurality of different podcasts). The model-generated script 206 may have been generated by generating a semantic understanding of the document 202, then generating a multi-turn dialogue discussing the content of the document 202 based on the semantic understanding. The model -generated script 206 can be descriptive of a podcast script that includes a host that poses questions and a technical expert that answers the questions. The answers to the questions can be generated based on the semantic understanding of the document 202.

[0058] The evaluation model 212 can then process the model -generated script 206 to generate an evaluation model output 214. The evaluation model 212 can include a machine- learned model trained to detect and/or annotate one or more potential errors, which may include factuality errors, speaker selection errors, terminology errors, sensitivity errors, and/or other potential errors. The evaluation model 212 may include one or more natural language processing models, one or more detection models, and/or one or more classification models. The evaluation model output 214 can include script annotations, instructions for performing corrections, and/or a revision prompt for prompting the generative model 204 for the next generation instance.

[0059] The generative model 204 can then process the model -generated script 206, the document 202, and/or the evaluation model output 214 to generate a revised script 216. The revised script 216 may have been generated by generating a semantic understanding of the document 202, then generating a multi-turn dialogue discussing the content of the document 202 based on the semantic understanding. The revised script 216 can be descriptive of a podcast script that includes a host that poses questions and a technical expert that answers the questions. The answers to the questions can be generated based on the semantic understanding of the document 202. The revised script 216 may be a top-down newly generated script and/or may be an augmented version of the model -generated script 206, which may have been augmented based on the evaluation model output 214. The generative model 204 may leverage the model-generated script 206 and the evaluation model output 214 for performing guided predictions during the revised script 216 generation.

[0060] A text-to-speech model 208 can then process the revised script 216 to generate audio data 210. The audio data 210 can be descriptive of the revised script 216 being spoken by two or more different speaker voices. The audio data 210 may be configured to be played with one or more audio player interfaces. The text-to-speech model 208 may include one or more variational autoencoders and/or one or more transformer models.

[0061] In some implementations, the model-generated script 206 and/or the revised script 216 may be generated based on a prompt 218 input into the generative model 204. The prompt 218 can include lines of text, lines of code, tuned weights or parameters, and/or other data. The prompt 218 may include instructions to generate a script based on one or more criteria. The prompt 218 may request for twenty or more dialogue turns. The prompt 218 may include criteria for pacing, tone, number of speakers, role of speakers, level of detail, timing, and/or other attributes.

[0062] Figure 3 depicts a flow chart diagram of an example method to perform according to example embodiments of the present disclosure. Although Figure 3 depicts steps performed in a particular order for purposes of illustration and discussion, the methods of the present disclosure are not limited to the particularly illustrated order or arrangement. The various steps of the method 300 can be omitted, rearranged, combined, and/or adapted in various ways without deviating from the scope of the present disclosure.

[0063] At 302, a computing system can obtain a document. The document can include an academic paper, experimental results, a presentation transcript, a slide deck, an excel sheet, a report, a slip opinion, a court transcript, and/or other documents. The document can be obtained from a user via a user interface. The user interface can include a graphical user interface with an upload interface. The document may be in a portable document format, an image file, a code file, and/or other file type. The document may include text data, image data, audio data, latent encoding data, statistical data, multimodal data, and/or other data.

[0064] At 304, the computing system can process the document with a generative language model to generate a model-generated script. The model-generated script can include a plurality of interleaved dialogue turns between two or more speakers. The model -generated script may be associated with a podcast script. In some implementations, the generative language model may have been trained on a plurality of real world training examples. The plurality of real world training examples may include long-form podcast examples.

Alternatively and/or additionally, the generative language model may be tuned on podcast transcripts. The generative language model may include a text encoder, an image encoder, an audio encoder, and/or a decoder. In some implementations, the generative language model may include one or more transformer models. The generative language model may include a vision language model for processing text and images of a document. The document may be processed with a prompt that conditions the generative language model to generate the podcast script with one or more criteria.

[0065] In some implementations, processing the document with the generative language model to generate the model-generated script can include generating semantic understanding data based on content and structure of the academic paper, generating a first set of questions, generating a first set of responses responsive to the first set of questions based on the semantic understanding data, and interweaving the first set of questions and the first set of responses to generate a chain of dialogue. The first set of questions can be generated based on determining the academic paper is associated with a particular information type based on the semantic understanding data.

[0066] The computing system can evaluate, with a machine-learned evaluation model, the model -generated script to generate an evaluation model output. The machine- learned evaluation model may process the model -generated transcript to determine whether the model -generated script includes inaccuracies, improper tone, improper style, vulgarity, and/or other improper attributes. The evaluation model output may include an annotation of the model -generated script. Alternatively and/or additionally, the evaluation model output may include a revision prompt for revising the model-generated script.

[0067] The computing system can process the document, the model -generated script, and the evaluation model output with the generative language model to generate a revised script. The revised script can include a first set of dialogue associated with a first speaker and a second set of dialogue associated with a second speaker. In some implementations, the first set of dialogue can be associated with a host role. The second set of dialogue can be associated with a technical expert role. Additionally and/or alternatively, the first set of dialogue can include questions. The second set of dialogue can include responses to the questions. The responses can be generated based on content of the academic paper. In some implementations, the revised script can include a second plurality of interleaved dialogue turns between the two or more speakers. The revised script can include a second modelgenerated script that differs from the model -generated script based on the generative language model being conditioned on the evaluation model output.

[0068] At 306, the computing system can process the model-generated script (and/or a revised script) with a text-to-speech model to generate audio data. The audio data can be descriptive of the model -generated script (and/or the revised script) being output in audio form by a first speaker voice and a second speaker voice. The audio data may emulate two podcast hosts reading the script. The first speaker voice and the second speaker voice may include different tones, paces, and/or pitches.

[0069] At 308, the computing system can provide the audio data for playback. The audio data may be provided for playback within the user interface. The audio data may be playable via a music player, a podcast player, and/or other audio playback interface.

[0070] In some implementations, the computing system can process the revised script and the audio data to determine the second speaker voice speaks a dialogue line associated with the first speaker. The computing system can then generate an additional text-to-speech prompt in response to determining the second speaker voice speaks the dialogue line associated with the first speaker. The computing system can then process the revised script and the additional text-to-speech prompt with the text-to-speech model to generate second audio data.

[0071] Figure 4 depicts an illustration of an example upload interface 400 according to example embodiments of the present disclosure. In particular, the upload interface 400 can include a plurality of options for uploading a document for paper-to-podcast conversion. The plurality of options can include an upload element 402, which may be selectable to open an interface to select documents saved locally on a user computing device. Additionally and/or alternatively, a user may be provided with an option to view a library 404 of documents, which may include a library of documents associated with the user and/or a set of users. The user may select one or more of the documents in the library 404 for paper-to-podcast conversion. The plurality of options may include a search function 406 for searching one or more databases with one or more search engines for documents from a particular author, from a particular entity, about a particular topic, and/or within a certain time period. The search can be performed based on a search query and/or one or more preference selections. The search results may then be selected to perform the paper-to-podcast conversion. Once one or more documents are uploaded and/or selected, the generate user interface element 408 can be selected to trigger the performance of the script generation.

[0072] Figure 5 depicts an illustration of an example selection interface 500 according to example embodiments of the present disclosure. In particular, the selection interface 500 can provide search results 502 for display with options to select checkboxes 504 for selecting documents for the document-to-dialogue conversion. Once documents are selected, the user can select to add the documents and/or cancel selections via one or more user interface elements 506. The search results 502 can include a document title along with a source of the document (e.g., a URL or computing device drive location). The checkboxes 504 can be selectable to select and/or deselect search results 502 for inclusion.

[0073] Figure 6 A depicts an illustration of an example library interface 600 according to example embodiments of the present disclosure. In particular, the library interface 600 can display the different model-generated audio files that were generated via the document-to- dialogue audio conversion. The library interface 600 can include a personal library panel 602 for audio files generated based on user inputs. Additionally and/or alternatively, the library interface 600 can include a public library panel 606 for audio files generated based on other user inputs. The first audio file 604 from the personal library panel 602 was generated based on a plurality of documents (3 sources). The second audio file 608 from the public library panel 606 was generated based on a paper written by John Doe, Jane Smith, and Joseph White. The third audio file 610 from the public library panel 606 was generated based on a plurality of documents (6 sources). The library interface 600 can include a filter user interface element 612 for filtering the audio files based on different attributes (e.g., generation date, paper publication date, topic, number of speakers, length, paper type of the original paper, number of papers associated with audio file, etc.).

[0074] Figure 6B depicts an illustration of an example playback interface 614 according to example embodiments of the present disclosure. In particular, the playback interface 614 may be provided for display in response to a particular audio file being selected for playback (e.g., when the first audio file 604 is selected for playback). The playback interface 614 may include an interactive progress bar, a pause/play button, a rewind button, a fast forward button, one or more skip buttons, a share user interface element, a transcript display user interface element (e.g., displaying the script associated with the audio file), a like user interface element, a dislike user interface element, an exit user interface element, a rate user interface element, one or more augmentation user interface elements, and/or other user interface features.

[0075] Figure 7 depicts a flow chart diagram of an example method to perform according to example embodiments of the present disclosure. Although Figure 7 depicts steps performed in a particular order for purposes of illustration and discussion, the methods of the present disclosure are not limited to the particularly illustrated order or arrangement. The various steps of the method 700 can be omitted, rearranged, combined, and/or adapted in various ways without deviating from the scope of the present disclosure.

[0076] At 702, a computing system can obtain a document. The document can include an academic paper. The academic paper can be a research paper with an abstract, an introduction, a related works section, a method section, an experiments section, and/or a conclusion section. The document can include structured text, media content, and/or tables. [0077] At 704, the computing system can process the document with a generative language model to generate a model-generated script. The model-generated script can include a plurality of interleaved dialogue turns between two or more speakers. In some implementations, the generative language model can include an autoregressive language model.

[0078] At 706, the computing system can evaluate, with a machine-learned evaluation model, the model -generated script to generate an evaluation model output. The machine- learned evaluation model can include a natural language processing model that includes a plurality of classifiers and a prompt generation model.

[0079] In some implementations, evaluating, with the machine-learned evaluation model, the model -generated script to generate the evaluation model output can include processing the model -generated script and the document with the machine-learned evaluation model to determine the model-generated script includes one or more factual inaccuracies and generating, with the machine-learned evaluation model, the evaluation model output that includes one or more facts from the document.

[0080] Alternatively and/or additionally, evaluating, with the machine-learned evaluation model, the model-generated script to generate the evaluation model output can include processing the model -generated script and the document with the machine-learned evaluation model to determine the model -generated script includes a first person perspective when referring to the document and generating, with the machine-learned evaluation model, the evaluation model output that includes instructions to include dialogue from a perspective of a reader.

[0081] At 708, the computing system can process the document, the model -generated script, and the evaluation model output with the generative language model to generate a revised script. The revised script may be an augmented version of the model -generated script that is augmented based on the evaluation model output.

[0082] In some implementations, the computing system can determine the academic paper is older than a threshold age. The computing system can obtain an age-disclaimer prompt in response to determining the academic paper is older than the threshold age. The computing system can then process the age-disclaimer prompt, the revised script, the modelgenerated evaluation output, and the document with the generative language model to generate an augmented script. The augmented script can include an augmented version of the revised script. In some implementations, the revised script can be augmented to include one or more disclaimers on a year of publication of the academic paper.

[0083] At 710, the computing system can process the revised script with a text-to- speech model to generate audio data. The audio data can be descriptive of the revised script being output in audio form by a first speaker voice and a second speaker voice. The audio data may include volume and pitch regularization for the two or more speakers.

[0084] Figure 8 depicts a flow chart diagram of an example method to perform according to example embodiments of the present disclosure. Although Figure 8 depicts steps performed in a particular order for purposes of illustration and discussion, the methods of the present disclosure are not limited to the particularly illustrated order or arrangement. The various steps of the method 800 can be omitted, rearranged, combined, and/or adapted in various ways without deviating from the scope of the present disclosure.

[0085] At 802, a computing system can obtain a document. The document can include an academic paper, a report, a research paper, experimental results, a transcript, and/or other documents. The document can be obtained based on pulling the document from one or more databases in response to a user selection, a URL input, and/or a user input upload.

[0086] At 804, the computing system can process the document with a generative language model to generate a model-generated script. The model-generated script can include a plurality of interleaved dialogue turns between two or more speakers. The two or more speakers may each be assigned a role in the dialogue, which may include an interviewer and a technical expert. The interviewer may ask questions, provide introductions, and make transitions. The technical expert may answer the questions based on the contents of the document, may perform monologues on the contents of the document, and/or may provide rebuttals.

[0087] At 806, the computing system can process the model-generated script with a machine-learned evaluation model to generate an evaluation model output. The machine- learned evaluation model may be trained to detect instances of facts being provided without factual grounding from the document. Additionally and/or alternatively, the machine-learned evaluation model may have been trained to detect when one or more of the dialogue lines include a perspective that insinuates authorship or contribution by one of the speakers.

[0088] At 808, the computing system can generate a revision prompt based on the evaluation model output. The revision prompt may include one or more lines of text indicating how to augment the model-generated script to fix a determined error. Alternatively and/or additionally, the revision prompt may include an outline of the model -generated script with a correction performed.

[0089] At 810, the computing system can process the document, the revision prompt, and the evaluation model output with the generative language model to generate a revised script. The revised script may be a top-down new generation and/or may be an augmented version of the model-generated script. The revised script may include twenty or more turns of dialogue.

[0090] At 812, the computing system can process the revised script with a text-to- speech model to generate audio data. In some implementations, the audio data can be descriptive of the revised script being output in audio form by a first speaker voice and a second speaker voice. The text-to-speech model can include a text-to-speech autoencoder. The text-to-speech model can include a transformer model.

[0091] In some implementations, the computing system can obtain a second document. The second document can include a second academic paper. The computing system can concatenate the document and the second document to generate a combined document. The computing system can process the combined document with the generative language model to generate a multi-document script.

[0092] The systems and methods disclosed herein may be utilized for automatic generation of natural -sounding audio content from a variety of input data. The proposed systems and methods address several technical challenges observed in previous attempts to apply machine learning to this task. In particular, when applied to task of automatic audio content generation, traditional large models such as large language models (LLMs) or similar often produced audio content that sounded unnatural due to issues such as preachy tones, excessive flattery, awkward transitions, monotone delivery, and/or limited conversation length. These drawbacks can make the generated dialogues feel less engaging and realistic. [0093] In contrast, the systems and methods disclosed herein can generate audio content that sounds more natural and engaging. The improved result can be achieved through a series of processes that include obtaining and processing large sets of context data, generating transcripts, and converting these transcripts into audio formats that users can listen to. For instance, the system can handle inputs ranging from text to multimodal data, such as videos, which are then converted into a textual format for further processing. Specifically, one aspect of the present disclosure can relate to the generation of long-form content by creating structured outlines and detailed sections through hierarchical processing. Another aspect of the present disclosure can improve user interaction with the generated audio content by incorporating real-time feedback mechanisms. Users can influence the content dynamically during playback through commands or questions, which the system processes to update the audio content accordingly. The real-time adaptability can enhance user experience by making audio content more responsive and personalized.

[0094] Thus, example aspects of the present disclosure can advance the field of automated audio content generation by introducing methods that produce more natural, engaging, and interactive audio experiences. The systems and methods disclosed herein can be particularly advantageous in applications such as virtual learning environments, interactive podcasts, and automated news reporting, where the quality of audio content significantly impacts user engagement and information retention. [0095] More particularly, an example computing system for generating audio content can first obtain a large corpus of context data. In some implementations, the context data can encompass a wide variety of types, forms, or sources of content, which may include textual documents, audio files, videos, and/or live feeds from different domains such as news, academia, or entertainment. This data can be automatically selected and/or can be curated by a human user to focus on specific topics of interest. To provide an example, a user may compile a selection of scholarly articles, expert lectures, and recent news clips all pertaining to quantum computing to serve as the input for generating contextually-relevant audio content (e.g., a podcast or tutorial relating to quantum computing). The system can allow users to tailor the computing system to produce customized content that meets their specific informational needs or preferences (or those of their intended audience, which may differ from the specific user controlling the system).

[0096] Next, the computing system can process the context data using one or more machine-learned sequence processing models to generate a transcript for the intended audio content. The transcript can include descriptions of various topics extracted from the context data. The use of machine-learned models can allow for the accurate identification and extraction of relevant topics from a wide array of data types, including textual and multimodal inputs. For instance, in the context of generating a podcast episode from a collection of academic works about quantum computing, the system can effectively discern and summarize key topics from the articles to be included in the transcript.

[0097] The computing system can then generate audio content from the prepared transcript, wherein the audio content comprises speech that verbalizes the transcript effectively. The audio generation can include transforming the textual representations of the transcript into spoken words using advanced voice synthesis technologies. As one example, the system can utilize machine-learned voice models that are capable of producing naturalsounding speech, closely mimicking human intonation and pronunciation.

[0098] Once the audio content has been generated, the computing system can provide the audio content for playback to a user. For example, the system can stream the audio content directly to the user’s device or make it available for download, depending on the user’s preference and the application’s design. The system can also support multiple formats of audio files, making it compatible with a wide range of devices and media players.

[0099] In some implementations, the computing system can provide the generated audio content for playback through streaming via a real-time communication (RTC) framework (e.g., WebRTC). The method can allow the audio content to be streamed directly to users in real-time, facilitating immediate and seamless delivery of content as it is generated or updated. For example, in a live interactive podcast or a real-time educational webinar, the audio content can be streamed to listeners across different geographical locations without significant delays. The streaming can increase the likelihood that all participants receive the content simultaneously and can interact with it in a timely manner. The feature can be particularly advantageous in scenarios where immediate user interaction is desired, such as in live news broadcasting or during interactive learning sessions.

[0100] In some implementations, the corpus of content data may be a set of userspecific data that has been curated by the personalized content curation service. For example, a personalized content curation service can offer a dynamic and tailored experience by delivering a daily feed of articles, news stories, and other relevant content based on individual user preferences and interests. By analyzing user data with the user’s consent, the service can be able to curate a corpus of data that is highly customized and constantly updated to ensure that users receive the most pertinent and engaging information. The approach to content delivery can help users stay informed and connected with topics that matter most to them, enhancing their overall online experience.

[0101] According to one aspect of the present disclosure, the systems and methods can enable the user to pause and interact with the audio content during playback of the audio content. For example, in some implementations, the computing system can receive interaction data from the user (e.g., during playback of the original audio content). The interaction data can include additional conditioning inputs, with examples described in further detail below. The interaction data can be processed alongside at least a portion of the previously-generated transcript using one or more machine-learned sequence processing models. For instance, if a user requests clarification on a specific topic mentioned in the audio content, the system can receive this request as interaction data and use it to revise the relevant section of the transcript. The updated transcript can then be used to generate new audio content that reflects the user’s input. The updated audio content can be provided for playback. The feature can be particularly advantageous in educational settings where learners might need additional explanations or in interactive storytelling applications where listeners might want to explore different plot directions based on their choices.

[0102] In some implementations, the computing system can receive interaction data from the user during the playback of the audio content. For example, a user may provide feedback or ask a question while listening to a podcast or lecture. Upon receiving such interaction data, the system can pause the ongoing playback of the audio content. The pause can allow the system to process the received interaction data and update the transcript accordingly, thereby generating new audio content that addresses the user’s input. Once the updated audio content is ready, playback can resume, possibly starting from the point of interruption or from a new relevant section, thus providing a seamless and interactive listening experience.

[0103] In some implementations, the computing system can specifically process the portion of the transcript that temporally follows the interaction time associated with the received interaction data. For instance, if a user provides feedback or asks a question at a specific time during the audio playback, the system can identify this interaction time and focus on updating the subsequent sections of the transcript that come after this point. The approach can increase the likelihood that the updates are relevant and timely. The relevancy can enhance the user’s experience by directly addressing their immediate concerns or interests in the ongoing content.

[0104] In some implementations, the computing system can handle a variety of additional conditioning inputs received from users to revise the audio content. These inputs can include interjections, clarification requests, on-topic questions, as well as inputs steering the focus, exclusion of topics, target audience level, host persona, or the tone and format of the show. For example, if a user inputs a focus steering request such as “Talk more about Josephson junctions”, the system can adjust the remaining part of the transcript to concentrate specifically on that topic. Similarly, if a user requests a change in the tone of the show to resemble a short, engaging presentation rather than an academic lecture, the system can modify the delivery style of the content while keeping the core information the same. Each type of conditioning input can lead to modifications in the length and/or content of the transcript. The adaptive feature can allow for a dynamic and interactive audio experience, allowing the user to interactively control the tone, substance, and/or style of the audio content.

[0105] In some implementations, the computing system can capture interaction data through user speech data obtained via a microphone. For instance, while listening to an audio presentation, a user may have questions or comments and can express these verbally. The system can then capture this spoken input directly through the microphone, process it to understand the user’s intent, and use this information to modify the ongoing audio content accordingly. The feature can allow for a hands-free interaction experience where users can engage with the content more naturally and conveniently. The engagement accessibility can enhance accessibility and user-friendliness. Such a capability can be particularly advantageous in scenarios where users are multitasking or when the technology is being used in environments like vehicles or while exercising, where manual text input is impractical. [0106] Another aspect of the present disclosure can be directed to techniques for handling sets of content data that have large or “massive” size. As one example, in some implementations, the computing system can process large volumes of context data by first dividing the context data into a plurality of chunks. Each chunk can then be individually processed by one or more machine-learned sequence processing models to generate reduced- data-size content chunks. For example, a document or a long video can be segmented into smaller, manageable parts, each part focusing on specific sections or topics. The smaller chunks can be easier to handle and analyze. The smaller chunks can enhance the efficiency and accuracy of data processing. Subsequently, the reduced-data-size content chunks can be further processed collectively to generate a transcript. The method can increase the likelihood that the final transcript is a coherent and accurate representation of the original context data, facilitating the generation of detailed and precise audio content. The approach can be particularly advantageous in scenarios including complex and voluminous data sets, such as multi-modal data, extensive technical documents, and/or long-form textual content such as novels or textbooks.

[0107] In some implementations, the generation of the transcript can include a step where at least one video contained in the context data is first transformed into a textual format before further processing to generate the transcript. For example, a video of a lecture or a news broadcast can be converted into text using advanced speech recognition technologies that accurately transcribe spoken words into written form and/or using a vision language model or other multi-modal model to create a textual summary of the video content. The conversion may be referred to as “semantic compression.” The conversion can allow the system to integrate and process video content alongside textual data. The integration can increase the likelihood that all relevant information, regardless of its original format, is considered in the generation of the transcript. The capability can be particularly advantageous in scenarios where important information is delivered in multimedia formats, which would otherwise be unwieldy or computationally expensive to jointly process in its native modality with all other context data.

[0108] Another aspect of the present disclosure can be directed to techniques for improving the ability to output audio content of significant length, while retaining semantic consistency and structure. As one example, in some implementations, the computing system can hierarchically generate the transcript by first creating a structured outline from the context data using one or more machine-learned sequence processing models. The outline can include a plurality of sections, each representing a distinct topic or segment of the overall content. For example, in the context of an educational series, the outline may include sections for introduction, key concepts, case studies, and conclusion. Following the creation of the outline, the system can iteratively process each section of the context data to generate a detailed transcript for each respective section. The hierarchical approach can increase the likelihood that the final transcript is well-organized and maintains logical coherence throughout the audio content. The transcript flow can be particularly advantageous for complex or lengthy informational material that requires clear segmentation to enhance listener comprehension and/or engagement.

[0109] Additional aspects of the present disclosure can be directed to techniques for creating audio content that is engaging and natural sounding from the perspective of the human listener. For example, in some implementations, the computing system can generate an initial transcript from the context data using one or more machine-learned sequence processing models. The initial transcript can serve as a preliminary version of the audio content. The system can perform one or more critic-driven rewrite loops to refine and edit the initial transcript, thereby producing an edited transcript. For example, the rewrite loops can include analyzing the initial transcript for any inaccuracies, inconsistencies, or areas that lack clarity or natural flow in the dialogue. The system can then perform the necessary adjustments to enhance the quality and accuracy of the content. The iterative process can increase the likelihood that the final transcript is not only accurate but also engaging and natural-sounding.

[0110] In some implementations, a machine-learned sequence processing model can be used as a computer-implemented critic to improve the quality of audio content through a critic-driven rewrite loop. For example, the sequence processing model can be equipped with specific critique instructions that detail the aspect(s) of the audio content to be critiqued, the desired format of the critique(s), and other relevant guidance to refine the critique process. The model can process the initial transcript according to these instructions and outputs critique(s), potentially in a natural language format. The critique(s) can identify areas for improvement and/or provide specific suggestions for improvement. Based on these critique(s), either the same or a different sequence processing model can then reprocess the initial transcript together with the critique(s) to produce a revised transcript. The revised transcript can be specifically rewritten to address the identified critiquejs, thereby improving the naturalness, accuracy, and/or relevance of the final audio content. The iterative process of critique and revision can increase the likelihood that the generated audio is engaging and natural sounding from the human’s perspective, or otherwise meets expectations of quality or accuracy.

[0111] As one example, in some implementations, the computing system can enhance the naturalness of the audio content by incorporating an increased amount of disfluencies into the transcript during the critic-driven rewrite loops. Disfluencies, such as slight hesitations, repetitions, or filler words like “um” and “ah,” can be common in natural speech and can make synthesized audio content sound more realistic and relatable. For example, when generating a podcast or an interactive dialogue for a virtual assistant, intentionally adding these disfluencies can prevent the audio from sounding too polished or mechanical.

[0112] As another example, critic-driven rewrite loops can be performed to enhance the accuracy or “groundedness” of the transcript relative to the source context material. As one example, in some implementations, the computing system can further refine the audio content by processing both the context data and the initial transcript with one or more machine-learned sequence processing models during the critic-driven rewrite loops. The process can aim to generate an edited transcript that exhibits increased grounding relative to the context data. For example, if the initial transcript derived from a detailed technical manual lacks certain nuances or context-specific terminologies, the rewrite loops can reintegrate these elements. The loop process flow can enhance the transcript’s fidelity to the source material. The approach can increase the likelihood that the final audio content not only sounds natural but also maintains a high level of accuracy and relevance.

[0113] As another example, in some implementations, the computing system can enhance the accuracy and clarity of audio content through a re-write loop that identifies and corrects verbalization errors in the ultimate audio content. For example, the system can initially generate audio content from the initial transcript. The initial audio can then be analyzed by one or more machine-learned sequence processing models. The models may be multi-modal models that are adept at detecting verbalization errors such as mispronunciations, grammatical inconsistencies, or unnatural phrasing. As another example, the audio content can be turned back into an additional transcript by a speech-to-text tool. The additional transcript can be compared to the initial transcript to identify any areas of divergence, which would indicate that the audio content does not successfully verbalize the initial transcript (as the additional transcript generated from the audio does not match the intended content). Once any verbalization errors are identified, the system can re-write the initial transcript, removing or replacing the erroneous portions. For example, if the initial audio incorrectly pronounces a technical term, the system can correct these issues in the transcript. The corrections can increase the likelihood that the subsequent audio output is both accurate and correctly verbalized.

[0114] As another example, in some implementations, the computing system can personalize the audio content by integrating an audience persona description into the critic- driven re-write loops. The process can include the computing system using the initial transcript along with a detailed description of the audience persona, which could include preferences, specific interests, and/or other information about the intended audience, to tailor the content. For example, the one or more machine-learned sequence processing models can process the initial transcript and the audience persona description to generate an edited transcript that is specifically conditioned on and personalized towards the audience persona. For example, if the audience persona indicates a preference for concise and straightforward explanations, the edited transcript may be adjusted to simplify explanations and avoid jargon. The personalized approach can increase the likelihood that the audio content is more engaging and relevant to the listener.

[0115] In some implementations, the computing system can generate the audio content from the transcript by processing it through a machine-learned voice model. The model can be trained to convert text into natural -sounding speech which mimics human intonation and rhythm. In some implementations, the voice model can be conditioned with data that indicates the particular voice profile or style for one or more of the synthetic speakers contained in the audio content. For instance, the voice model can take a finalized transcript and produce audio that sounds like a human newsreader for a news broadcast or a narrator for an audiobook. In other implementations, standard text-to-speech tools can be used to generate the audio content.

[0116] In some implementations, the computing system can additionally generate video content (or other supporting or correlated content) that is either temporally or semantically synchronized with the audio content. The result may mean that the computing system can create visuals that align with the timing and context of the spoken words. For example, the system can output a video that depicts synthetic persons verbalizing the corresponding audio content. The generation of additional supporting or correlated content can enhance the overall multimedia experience. For example, during a news broadcast, as the audio content describes a specific event, corresponding video footage or graphical representations of the event can be displayed. Similarly, in an educational setting, as a concept is explained audibly, relevant diagrams or animations can be shown to aid in comprehension.

[0117] In some implementations, the computing system can include a user interface designed to enhance the playback experience of the audio content by providing attribution or citation elements. For example, these elements can be displayed to the user during the audio playback and can be temporally synchronized with the content being played. For example, when a particular scientific study or historical event is mentioned in the audio, the user interface can simultaneously display a citation or attribution linked to that reference, providing immediate access to the source material. The feature can be advantageous in settings where verifying the accuracy and source of information is important. The result can add a layer of transparency to the content, allowing users to explore the origins of the information in real-time as they listen.

[0118] In some implementations, the technology may be capable of generating a diverse range of audio content types to suit various user needs and preferences. For example, the audio content produced can include dialogs, which are conversational pieces between two or more parties, podcasts that cover myriad topics in episodic formats, educational content tailored for learning and development, sports commentary providing real-time or summarized insights into sporting events, news commentary that discusses current events, and/or summaries of recent events or other fresh content. The versatility can make the technology suitable for a wide array of applications, from entertainment and education to sports broadcasting and news reporting.

[0119] The systems and methods of the present disclosure provide a number of technical effects and benefits. As one example, the proposed technology enhances the processing efficiency and accuracy of generating audio content from diverse data inputs, including textual and video data. By utilizing advanced machine-learned sequence processing models, the system can transform large volumes of context data into coherent audio outputs. The technical feature can address the technical challenge of managing and synthesizing vast and varied data types. Another technical benefit relates to the use of critic-driven rewrite loops to refine the generated audio content. The loops can allow the system to iteratively improve the audio output by identifying and correcting verbalization errors and/or enhancing the content’s alignment with user preferences.

[0120] Another significant technical aspect of the proposed technology is its ability to generate audio content that is synchronized with and/or responsive to real-time user interactions. The dynamic interaction capability can be facilitated by the system’s real-time processing of user inputs, such as questions or commands, to modify the audio content on- the-fly. The feature can leverage real-time data handling, such as the use of a RTC framework. In particular, the use of an RTC can enable instantaneous streaming and interaction capabilities. RTC can facilitate the direct and immediate transmission of audio content over the internet, allowing users to receive and interact with the content with minimal latency. The reduced latency can be particularly advantageous as it enables real-time feedback. Moreover, RTC can support a range of data types, including audio and video, which allows the system to synchronize multiple media streams effectively.

[0121] Furthermore, allowing real-time interactive inputs from the user significantly reduces computational expenditure by enabling the system to focus processing resources on specific segments of the transcript that require modification, rather than rewriting the entire transcript. When a user provides an input, such as a question or a command, the system can identify the relevant portion of the ongoing transcript that needs adjustment and processes only that segment. The targeted approach can avoid the unnecessary computational load and resource usage that would be included in reprocessing the entire audio content. Additionally, the method can enhance efficiency by reducing the latency and processing time, as the system can quickly adapt the content based on real-time feedback without the need to queue changes for the entire document.

[0122] As another example technical benefit, the hierarchical processing of context data significantly can reduce computational expenditure by breaking down large datasets into manageable chunks before processing. The method can be particularly effective due to the computational complexity of attention mechanisms (e.g., which are commonly used in sequence processing models) scales quadratically with the size of the context being processed. By dividing the context data into smaller segments, the system can limit the size of the data each processing unit must handle at any one time, thereby reducing the number of computational operations required. This not only speeds up the processing time but also enhances the efficiency of the system, reducing the consumption of computational resources. [0123] Due to the inherent flexibility of computing systems, a variety of device and system configurations can be implemented to facilitate the generation and delivery of audio content. For instance, one example arrangement can include generating audio content on a server, which is then streamed to a user device through a real-time communication (RTC) framework. In this setup, user inputs can be captured at the user device and streamed back to the server, allowing the server to update the audio content based on this feedback.

Alternatively and/or additionally, configurations where all functionalities are performed directly on the user’s device, known as “on-device” configurations, are also feasible. The approach can benefit from potentially faster response times and enhanced privacy, as data does not need to be transmitted over a network. Other configurations of functionality can be possible as well.

[0124] Figure 9 depicts a flowchart of a method 900 for training one or more machine-learned models according to aspects of the present disclosure. For instance, an example machine-learned model can include a generative language model, a machine-learned model, and/or a text-to-speech model. The machine-learned models can encoders, decoders, classifiers, self-attention models, feed-forward models, convolutional models, transformer models, recurrent models, detection models, segmentation models, and/or other models. [0125] One or more portion(s) of example method 900 can be implemented by a computing system that includes one or more computing devices such as, for example, computing systems described with reference to the other figures. Each respective portion of example method 900 can be performed by any (or any combination) of one or more computing devices. Moreover, one or more portion(s) of example method 900 can be implemented on the hardware components of the device(s) described herein, for example, to train one or more systems or models. Figure 9 depicts elements performed in a particular order for purposes of illustration and discussion. Those of ordinary skill in the art, using the disclosures provided herein, will understand that the elements of any of the methods discussed herein can be adapted, rearranged, expanded, omitted, combined, or modified in various ways without deviating from the scope of the present disclosure. Figure 9 is described with reference to elements/terms described with respect to other systems and figures for exemplary illustrated purposes and is not meant to be limiting. One or more portions of example method 900 can be performed additionally, or alternatively, by other systems. [0126] At 902, example method 900 can include obtaining a training instance. A set of training data can include a plurality of training instances divided between multiple datasets (e.g., a training dataset, a validation dataset, or testing dataset). A training instance can be labeled or unlabeled. Although referred to in example method 900 as a “training” instance, it is to be understood that runtime inferences can form training instances when a model is trained using an evaluation of the model’s performance on that runtime instance (e.g., online training/learning). Example data types for the training instance and various tasks associated therewith are described throughout the present disclosure.

[0127] At 904, example method 900 can include processing, using one or more machine-learned models, the training instance to generate an output. The output can be directly obtained from the one or more machine-learned models or can be a downstream result of a chain of processing operations that includes an output of the one or more machine- learned models.

[0128] At 906, example method 900 can include receiving an evaluation signal associated with the output. The evaluation signal can be obtained using a loss function. Various determinations of loss can be used, such as mean squared error, likelihood loss, cross entropy loss, hinge loss, contrastive loss, or various other loss functions. The evaluation signal can be computed using known ground-truth labels (e.g., supervised learning), predicted or estimated labels (e.g., semi- or self-supervised learning), or without labels (e.g., unsupervised learning). The evaluation signal can be a reward (e.g., for reinforcement learning). The reward can be computed using a machine-learned reward model configured to generate rewards based on output(s) received. The reward can be computed using feedback data describing human feedback on the output(s).

[0129] At 908, example method 900 can include updating the machine-learned model using the evaluation signal. For example, values for parameters of the machine-learned model(s) can be learned, in some embodiments, using various training or learning techniques, such as, for example, backwards propagation. For example, the evaluation signal can be backpropagated from the output (or another source of the evaluation signal) through the machine-learned model(s) to update one or more parameters of the model(s) (e.g., based on a gradient of the evaluation signal with respect to the parameter value(s)). For example, system(s) containing one or more machine-learned models can be trained in an end-to-end manner. Gradient descent techniques can be used to iteratively update the parameters over a number of training iterations. In some implementations, performing backwards propagation of errors can include performing truncated backpropagation through time. Example method 900 can include implementing a number of generalization techniques (e.g., weight decays, dropouts, etc.) to improve the generalization capability of the models being trained.

[0130] In some implementations, example method 900 can be implemented for training a machine-learned model from an initialized state to a fully trained state (e.g., when the model exhibits a desired performance profile, such as based on accuracy, precision, recall, etc.).

[0131] In some implementations, example method 900 can be implemented for particular stages of a training procedure. For instance, in some implementations, example method 900 can be implemented for pre-training a machine-learned model. Pre-training can include, for instance, large-scale training over potentially noisy data to achieve a broad base of performance levels across a variety of tasks/data types. In some implementations, example method 900 can be implemented for fine-tuning a machine-learned model. Fine-tuning can include, for instance, smaller-scale training on higher-quality (e.g., labeled, curated, etc.) data. Fine-tuning can affect all or a portion of the parameters of a machine-learned model. For example, various portions of the machine-learned model can be “frozen” for certain training stages. For example, parameters associated with an embedding space can be “frozen” during fine-tuning (e.g., to retain information learned from a broader domain(s) than present in the fine-tuning dataset(s)). An example fine-tuning approach includes reinforcement learning. Reinforcement learning can be based on user feedback on model performance during use.

[0132] Figure 10 is a block diagram of an example processing flow for using machine-learned model(s) 1 to process input(s) 2 to generate output(s) 3.

[0133] Machine-learned model(s) 1 can be or include one or multiple machine- learned models or model components. Example machine-learned models can include neural networks (e.g., deep neural networks). Example machine-learned models can include nonlinear models or linear models. Example machine-learned models can use other architectures in lieu of or in addition to neural networks. Example machine-learned models can include decision tree based models, support vector machines, hidden Markov models, Bayesian networks, linear regression models, k-means clustering models, etc.

[0134] Example neural networks can include feed-forward neural networks, recurrent neural networks (RNNs), including long short-term memory (LSTM) based recurrent neural networks, convolutional neural networks (CNNs), diffusion models, generative-adversarial networks, or other forms of neural networks. Example neural networks can be deep neural networks. Some example machine-learned models can leverage an attention mechanism such as self-attention. For example, some example machine-learned models can include multiheaded self-attention models.

[0135] Machine-learned model(s) 1 can include a single or multiple instances of the same model configured to operate on data from input(s) 2. Machine-learned model(s) 1 can include an ensemble of different models that can cooperatively interact to process data from input(s) 2. For example, machine-learned model(s) 1 can employ a mixture-of-experts structure. See, e.g., Zhou et al., Mixture-of-Experts with Expert Choice Routing, ARXIV:2202.09368V2 (Oct. 14, 2022).

[0136] Input(s) 2 can generally include or otherwise represent various types of data. Input(s) 2 can include one type or many different types of data. Output(s) 3 can be data of the same type(s) or of different types of data as compared to input(s) 2. Output(s) 3 can include one type or many different types of data.

[0137] Example data types for input(s) 2 or output(s) 3 include natural language text data, software code data (e.g., source code, object code, machine code, or any other form of computer-readable instructions or programming languages), machine code data (e.g., binary code, assembly code, or other forms of machine-readable instructions that can be executed directly by a computer's central processing unit), assembly code data (e.g., low-level programming languages that use symbolic representations of machine code instructions to program a processing unit), genetic data or other chemical or biochemical data, image data, audio data, audiovisual data, haptic data, biometric data, medical data, financial data, statistical data, geographical data, astronomical data, historical data, sensor data generally (e.g., digital or analog values, such as voltage or other absolute or relative level measurement values from a real or artificial input, such as from an audio sensor, light sensor, displacement sensor, etc.), and the like. Data can be raw or processed and can be in any format or schema. [0138] In multimodal inputs 2 or outputs 3, example combinations of data types include image data and audio data, image data and natural language data, natural language data and software code data, image data and biometric data, sensor data and medical data, etc. It is to be understood that any combination of data types in an input 2 or an output 3 can be present.

[0139] An example input 2 can include one or multiple data types, such as the example data types noted above. An example output 3 can include one or multiple data types, such as the example data types noted above. The data type(s) of input 2 can be the same as or different from the data type(s) of output 3. It is to be understood that the example data types noted above are provided for illustrative purposes only. Data types contemplated within the scope of the present disclosure are not limited to those examples noted above.

[0140] Figure 11 is a block diagram of an example implementation of an example machine-learned model configured to process sequences of information. For instance, an example implementation of machine-learned model(s) 1 can include machine-learned sequence processing model(s) 4. An example system can pass input(s) 2 to sequence processing model(s) 4. Sequence processing model(s) 4 can include one or more machine- learned components. Sequence processing model(s) 4 can process the data from input(s) 2 to obtain an input sequence 5. Input sequence 5 can include one or more input elements 5-1, 5- 2, . . . , 5-A , etc. obtained from input(s) 2. Sequence processing model 4 can process input sequence 5 using prediction layer(s) 6 to generate an output sequence 7. Output sequence 7 can include one or more output elements 7-1, 7-2, . . . , 7 -A, etc. generated based on input sequence 5. The system can generate output(s) 3 based on output sequence 7.

[0141] Sequence processing model(s) 4 can include one or multiple machine-learned model components configured to ingest, generate, or otherwise reason over sequences of information. For example, some example sequence processing models in the text domain are referred to as “Large Language Models,” or LLMs. See, e.g., PaLM 2 Technical Report, GOOGLE, https://ai.google/static/documents/palm2techreport.pdf (n.d.). Other example sequence processing models can operate in other domains, such as image domains, see, e.g., Dosovitskiy et al., An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale, ARXIV:2010.11929v2 (Jun. 3, 2021), audio domains, see, e.g., Agostinelli et al., MusicLM: Generating Music From Text, ARXIV:2301.1 1325V1 (Jan. 26, 2023), biochemical domains, see, e.g., Jumper et al., Highly accurate protein structure prediction with AlphaFold, 596 Nature 583 (Aug. 26, 2021), by way of example. Sequence processing model(s) 4 can process one or multiple types of data simultaneously. Sequence processing model(s) 4 can include relatively large models (e.g., more parameters, computationally expensive, etc.), relatively small models (e.g., fewer parameters, computationally lightweight, etc.), or both. [0142] In general, sequence processing model(s) 4 can obtain input sequence 5 using data from input(s) 2. For instance, input sequence 5 can include a representation of data from input(s) 2 in a format understood by sequence processing model(s) 4. One or more machine- learned components of sequence processing model(s) 4 can ingest the data from input(s) 2, parse the data into pieces compatible with the processing architectures of sequence processing model(s) 4 (e.g., via “tokenization”), and project the pieces into an input space associated with prediction layer(s) 6 (e.g., via “embedding”).

[0143] Sequence processing model(s) 4 can ingest the data from input(s) 2 and parse the data into a sequence of elements to obtain input sequence 5. For example, a portion of input data from input(s) 2 can be broken down into pieces that collectively represent the content of the portion of the input data. The pieces can provide the elements of the sequence. [0144] Elements 5-1, 5-2, . . . , 5-M can represent, in some cases, building blocks for capturing or expressing meaningful information in a particular data domain. For instance, the elements can describe “atomic units” across one or more domains. For example, for textual input source(s), the elements can correspond to groups of one or more words or sub-word components, such as sets of one or more characters.

[0145] For example, elements 5-1, 5-2, . . . , 5M can represent tokens obtained using a tokenizer. For instance, a tokenizer can process a given portion of an input source and output a series of tokens (e.g., corresponding to input elements 5-1, 5-2, . . . , 5-M) that represent the portion of the input source. Various approaches to tokenization can be used. For instance, textual input source(s) can be tokenized using a byte-pair encoding (BPE) technique. See, e.g., Kudo et al., SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing, PROCEEDINGS OF THE 2018 CONFERENCE ON EMPIRICAL METHODS IN NATURAL LANGUAGE PROCESSING (System Demonstrations), pages 66-71 (October 31-November 4, 2018), https://aclanthology.org/D18-2012.pdf. Image-based input source(s) can be tokenized by extracting and serializing patches from an image.

[0146] In general, arbitrary data types can be serialized and processed into input sequence 5. It is to be understood that element(s) 5-1, 5-2, . . . , 5-M depicted in Figure 11 can be the tokens or can be the embedded representations thereof.

[0147] Prediction layer(s) 6 can predict one or more output elements 7-1, 7-2, . . . , 7- N based on the input elements. Prediction layer(s) 6 can include one or more machine-learned model architectures, such as one or more layers of learned parameters that manipulate and transform the input(s) to extract higher-order meaning from, and relationships between, input element(s) 5-1, 5-2, . . . , 5-M. In this manner, for instance, example prediction layer(s) 6 can predict new output element(s) in view of the context provided by input sequence 5.

[0148] Prediction layer(s) 6 can evaluate associations between portions of input sequence 5 and a particular output element. These associations can inform a prediction of the likelihood that a particular output follows the input context. For example, consider the textual snippet, “The carpenter’s toolbox was small and heavy. It was full of .” Example prediction layer(s) 6 can identify that “It” refers back to “toolbox” by determining a relationship between the respective embeddings. Example prediction layer(s) 6 can also link “It” to the attributes of the toolbox, such as “small” and “heavy.” Based on these associations, prediction layer(s) 6 can, for instance, assign a higher probability to the word “nails” than to the word “sawdust.”

[0149] A transformer is an example architecture that can be used in prediction layer(s) 4. See, e.g., Vaswani et al., Attention Is All You Need, ARXIV: 1706.03762v7 (Aug. 2, 2023). A transformer is an example of a machine-learned model architecture that uses an attention mechanism to compute associations between items within a context window. The context window can include a sequence that contains input sequence 5 and potentially one or more output element(s) 7-1, 7-2, . . . , 7-N. A transformer block can include one or more attention layer(s) and one or more post-attention layer(s) (e.g., feedforward layer(s), such as a multi-layer perceptron).

[0150] Prediction layer(s) 6 can include other machine-learned model architectures in addition to or in lieu of transformer-based architectures. For example, recurrent neural networks (RNNs) and long short-term memory (LSTM) models can also be used, as well as convolutional neural networks (CNNs). In general, prediction layer(s) 6 can leverage various kinds of artificial neural networks that can understand or generate sequences of information. [0151] Output sequence 7 can include or otherwise represent the same or different data types as input sequence 5. For instance, input sequence 5 can represent textual data, and output sequence 7 can represent textual data. Input sequence 5 can represent image, audio, or audiovisual data, and output sequence 7 can represent textual data (e.g., describing the image, audio, or audiovisual data). It is to be understood that prediction layer(s) 6, and any other interstitial model components of sequence processing model(s) 4, can be configured to receive a variety of data types in input sequence(s) 5 and output a variety of data types in output sequence(s) 7.

[0152] Output sequence 7 can have various relationships to input sequence 5. Output sequence 7 can be a continuation of input sequence 5. Output sequence 7 can be complementary to input sequence 5. Output sequence 7 can translate, transform, augment, or otherwise modify input sequence 5. Output sequence 7 can answer, evaluate, confirm, or otherwise respond to input sequence 5. Output sequence 7 can implement (or describe instructions for implementing) an instruction provided via input sequence 5.

[0153] Output sequence 7 can be generated autoregressively. For instance, for some applications, an output of one or more prediction layer(s) 6 can be passed through one or more output layers (e.g., softmax layer) to obtain a probability distribution over an output vocabulary (e.g., a textual or symbolic vocabulary) conditioned on a set of input elements in a context window. In this manner, for instance, output sequence 7 can be autoregressively generated by sampling a likely next output element, adding that element to the context window, and re-generating the probability distribution based on the updated context window, and sampling a likely next output element, and so forth.

[0154] Output sequence 7 can also be generated non-autoregressively. For instance, multiple output elements of output sequence 7 can be predicted together without explicit sequential conditioning on each other. See, e.g., Saharia et al., Non-Autoregressive Machine Translation with Latent Alignments, ARXIV:2004.07437V3 (NOV. 16, 2020). [0155] Output sequence 7 can include one or multiple portions or elements. In an example content generation configuration, output sequence 7 can include multiple elements corresponding to multiple portions of a generated output sequence (e.g., a textual sentence, values of a discretized waveform, computer code, etc.). In an example classification configuration, output sequence 7 can include a single element associated with a classification output. For instance, an output “vocabulary” can include a set of classes into which an input sequence is to be classified. For instance, a vision transformer block can pass latent state information to a multilayer perceptron that outputs a likely class value associated with an input image.

[0156] Figure 12 is a block diagram of an example technique for populating an example input sequence 8. Input sequence 8 can include various functional elements that form part of the model infrastructure, such as an element 8-0 obtained from a task indicator 9 that signals to any model(s) that process input sequence 8 that a particular task is being performed (e.g., to help adapt a performance of the model(s) to that particular task). Input sequence 8 can include various data elements from different data modalities. For instance, an input modality 10-1 can include one modality of data. A data-to- sequence model 11-1 can process data from input modality 10-1 to project the data into a format compatible with input sequence 8 (e.g., one or more vectors dimensioned according to the dimensions of input sequence 8) to obtain elements 8-1, 8-2, 8-3. Another input modality 10-2 can include a different modality of data. A data-to-sequence model 11-2 can project data from input modality 10-2 into a format compatible with input sequence 8 to obtain elements 8-4, 8-5, 8- 6. Another input modality 10-3 can include yet another different modality of data. A data-to- sequence model 11-3 can project data from input modality 10-3 into a format compatible with input sequence 8 to obtain elements 8-7, 8-8, 8-9.

[0157] Input sequence 8 can be the same as or different from input sequence 5. Input sequence 8 can be a multimodal input sequence that contains elements that represent data from different modalities using a common dimensional representation. For instance, an embedding space can have P dimensions. Input sequence 8 can be configured to contain a plurality of elements that have /^J dimensions. In this manner, for instance, example implementations can facilitate information extraction and reasoning across diverse data modalities by projecting data into elements in the same embedding space for comparison, combination, or other computations therebetween.

[0158] For example, elements 8-0, . . . , 8-9 can indicate particular locations within a multidimensional embedding space. Some elements can map to a set of discrete locations in the embedding space. For instance, elements that correspond to discrete members of a predetermined vocabulary of tokens can map to discrete locations in the embedding space that are associated with those tokens. Other elements can be continuously distributed across the embedding space. For instance, some data types can be broken down into continuously defined portions (e.g., image patches) that can be described using continuously distributed locations within the embedding space.

[0159] In some implementations, the expressive power of the embedding space may not be limited to meanings associated with any particular set of tokens or other building blocks. For example, a continuous embedding space can encode a spectrum of high-order information. An individual piece of information (e.g., a token) can map to a particular point in that space: for instance, a token for the word “dog” can be projected to an embedded value that points to a particular location in the embedding space associated with canine-related information. Similarly, an image patch of an image of a dog on grass can also be projected into the embedding space. In some implementations, the projection of the image of the dog can be similar to the projection of the word “dog” while also having similarity to a projection of the word “grass,” while potentially being different from both. In some implementations, the projection of the image patch may not exactly align with any single projection of a single word. In some implementations, the projection of the image patch can align with a combination of the projections of the words “dog” and “grass.” In this manner, for instance, a high-order embedding space can encode information that can be independent of data modalities in which the information is expressed.

[0160] Task indicator 9 can include a model or model component configured to identify a task being performed and inject, into input sequence 8, an input value represented by element 8-0 that signals which task is being performed. For instance, the input value can be provided as a data type associated with an input modality and projected along with that input modality (e.g., the input value can be a textual task label that is embedded along with other textual data in the input; the input value can be a pixel-based representation of a task that is embedded along with other image data in the input; etc.). The input value can be provided as a data type that differs from or is at least independent from other input(s). For instance, the input value represented by element 8-0 can be learned within a continuous embedding space.

[0161] Input modalities 10-1, 10-2, and 10-3 can be associated with various different data types (e.g., as described above with respect to input(s) 2 and output(s) 3). [0162] Data-to-sequence models 11-1, 11-2, and 11-3 can be the same or different from each other. Data-to-sequence models 11-1, 11-2, and 11-3 can be adapted to each respective input modality 10-1, 10-2, and 10-3. For example, a textual data-to-sequence model can subdivide a portion of input text and project the subdivisions into element(s) in input sequence 8 (e.g., elements 8-1, 8-2, 8-3, etc.). An image data-to-sequence model can subdivide an input image and project the subdivisions into element(s) in input sequence 8 (e.g., elements 8-4, 8-5, 8-6, etc.). An arbitrary datatype data-to-sequence model can subdivide an input of that arbitrary datatype and project the subdivisions into element(s) in input sequence 8 (e.g., elements 8-7, 8-8, 8-9, etc.).

[0163] Data-to-sequence models 11-1, 11-2, and 11-3 can form part of machine- learned sequence processing model(s) 4. Data-to-sequence models 11-1, 11-2, and 11-3 can be jointly trained with or trained independently from machine-learned sequence processing model(s) 4. Data-to-sequence models 11-1, 11-2, and 11-3 can be trained end-to-end with machine-learned sequence processing model(s) 4.

[0164] Figure 13 is a block diagram of an example model development platform 12 that can facilitate creation, adaptation, and refinement of example machine-learned models (e.g., machine-learned model(s) 1, sequence processing model(s) 4, etc.). Model development platform 12 can provide a number of different toolkits that developer systems can employ in the development of new or adapted machine-learned models.

[0165] Model development platform 12 can provide one or more model libraries 13 containing building blocks for new models. Model libraries 13 can include one or more pretrained foundational models 13-1, which can provide a backbone of processing power across various tasks. Model libraries 13 can include one or more pre-trained expert models 13-2, which can be focused on performance in particular domains of expertise. Model libraries 13 can include various model primitives 13-3, which can provide low-level architectures or components (optionally pre-trained), which can be assembled in various arrangements as desired.

[0166] Model development platform 12 can receive selections of various model components 14. Model development platform 12 can pass selected model components 14 to a workbench 15 that combines selected model components 14 into a development model 16.

[0167] Workbench 15 can facilitate further refinement and adaptation of development model 16 by leveraging a number of different toolkits integrated with model development platform 12. For example, workbench 15 can facilitate alignment of the development model 16 with a desired performance profile on various tasks using a model alignment toolkit 17. [0168] Model alignment toolkit 17 can provide a number of tools for causing development model 16 to generate outputs aligned with desired behavioral characteristics. Alignment can include increasing an accuracy, precision, recall, etc. of model outputs. Alignment can include enforcing output styles, schema, or other preferential characteristics of model outputs. Alignment can be general or domain-specific. For instance, a pre-trained foundational model 13-1 can begin with an initial level of performance across multiple domains. Alignment of the pre-trained foundational model 13-1 can include improving a performance in a particular domain of information or tasks (e.g., even at the expense of performance in another domain of information or tasks).

[0169] Model alignment toolkit 17 can integrate one or more dataset(s) 17-1 for aligning development model 16. Curated dataset(s) 17-1 can include labeled or unlabeled training data. Dataset(s) 17-1 can be obtained from public domain datasets. Dataset(s) 17-1 can be obtained from private datasets associated with one or more developer system(s) for the alignment of bespoke machine-learned model(s) customized for private use-cases.

[0170] Pre-training pipelines 17-2 can include a machine-learned model training workflow configured to update development model 16 over large-scale, potentially noisy datasets. For example, pre-training can leverage unsupervised learning techniques (e.g., denoising, etc.) to process large numbers of training instances to update model parameters from an initialized state and achieve a desired baseline performance. Pre-training pipelines 17-2 can leverage unlabeled datasets in dataset(s) 17-1 to perform pre-training. Workbench 15 can implement a pre-training pipeline 17-2 to pre-train development model 16.

[0171] Fine-tuning pipelines 17-3 can include a machine-learned model training workflow configured to refine the model parameters of development model 16 with higher- quality data. Fine-tuning pipelines 17-3 can update development model 16 by conducting supervised training with labeled dataset(s) in dataset(s) 17-1. Fine-tuning pipelines 17-3 can update development model 16 by conducting reinforcement learning using reward signals from user feedback signals. Workbench 15 can implement a fine-tuning pipeline 17-3 to finetune development model 16.

[0172] Prompt libraries 17-4 can include sets of inputs configured to induce behavior aligned with desired performance criteria. Prompt libraries 17-4 can include few-shot prompts (e.g., inputs providing examples of desired model outputs for prepending to a desired runtime query), chain-of-thought prompts (e.g., inputs providing step-by-step reasoning within the exemplars to facilitate thorough reasoning by the model), and the like. [0173] Example prompts can be retrieved from an available repository of prompt libraries 17-4. Example prompts can be contributed by one or more developer systems using workbench 15.

[0174] In some implementations, pre-trained or fine-tuned models can achieve satisfactory performance without exemplars in the inputs. For instance, zero-shot prompts can include inputs that lack exemplars. Zero-shot prompts can be within a domain within a training dataset or outside of the training domain(s).

[0175] Prompt libraries 17-4 can include one or more prompt engineering tools.

Prompt engineering tools can provide workflows for retrieving or learning optimized prompt values. Prompt engineering tools can facilitate directly learning prompt values (e.g., input element values) based on one or more training iterations. Workbench 15 can implement prompt engineering tools in development model 16.

[0176] Prompt libraries 17-4 can include pipelines for prompt generation. For example, inputs can be generated using development model 16 itself or other machine- learned models. In this manner, for instance, a first model can process information about a task and output and input for a second model to process in order to perform a step of the task. The second model can be the same as or different from the first model. Workbench 15 can implement prompt generation pipelines in development model 16.

[0177] Prompt libraries 17-4 can include pipelines for context injection. For instance, a performance of development model 16 on a particular task can improve if provided with additional context for performing the task. Prompt libraries 17-4 can include software components configured to identify desired context, retrieve the context from an external source (e.g., a database, a sensor, etc.), and add the context to the input prompt. Workbench 15 can implement context injection pipelines in development model 16.

[0178] Although various training examples described herein with respect to model development platform 12 refer to “pre-training” and “fine-tuning,” it is to be understood that model alignment toolkit 17 can generally support a wide variety of training techniques adapted for training a wide variety of machine-learned models. Example training techniques can correspond to the example training method 900 described above.

[0179] Model development platform 12 can include a model plugin toolkit 18. Model plugin toolkit 18 can include a variety of tools configured for augmenting the functionality of a machine-learned model by integrating the machine-learned model with other systems, devices, and software components. For instance, a machine-learned model can use tools to increase performance quality where appropriate. For instance, deterministic tasks can be offloaded to dedicated tools in lieu of probabilistically performing the task with an increased risk of error. For instance, instead of autoregressively predicting the solution to a system of equations, a machine-learned model can recognize a tool to call for obtaining the solution and pass the system of equations to the appropriate tool. The tool can be a traditional system of equations solver that can operate deterministically to resolve the system of equations. The output of the tool can be returned in response to the original query. In this manner, tool use can allow some example models to focus on the strengths of machine-learned models — e.g., understanding an intent in an unstructured request for a task — while augmenting the performance of the model by offloading certain tasks to a more focused tool for rote application of deterministic algorithms to a well-defined problem.

[0180] Model plugin toolkit 18 can include validation tools 18-1. Validation tools 18- 1 can include tools that can parse and confirm output(s) of a machine-learned model.

Validation tools 18-1 can include engineered heuristics that establish certain thresholds applied to model outputs. For example, validation tools 18-1 can ground the outputs of machine-learned models to structured data sources (e.g., to mitigate “hallucinations”). [0181] Model plugin toolkit 18 can include tooling packages 18-2 for implementing one or more tools that can include scripts or other executable code that can be executed alongside development model 16. Tooling packages 18-2 can include one or more inputs configured to cause machine-learned model(s) to implement the tools (e.g., few-shot prompts that induce a model to output tool calls in the proper syntax, etc.). Tooling packages 18-2 can include, for instance, fine-tuning training data for training a model to use a tool.

[0182] Model plugin toolkit 18 can include interfaces for calling external application programming interfaces (APIs) 18-3. For instance, in addition to or in lieu of implementing tool calls or tool code directly with development model 16, development model 16 can be aligned to output instructions that initiate API calls to send or obtain data via external systems.

[0183] Model plugin toolkit 18 can integrate with prompt libraries 17-4 to build a catalog of available tools for use with development model 16. For instance, a model can receive, in an input, a catalog of available tools, and the model can generate an output that selects a tool from the available tools and initiates a tool call for using the tool.

[0184] Model development platform 12 can include a computational optimization toolkit 19 for optimizing a computational performance of development model 16. For instance, tools for model compression 19-1 can allow development model 16 to be reduced in size while maintaining a desired level of performance. For instance, model compression 19-1 can include quantization workflows, weight pruning and sparsification techniques, etc. Tools for hardware acceleration 19-2 can facilitate the configuration of the model storage and execution formats to operate optimally on different hardware resources. For instance, hardware acceleration 19-2 can include tools for optimally sharing models for distributed processing over multiple processing units for increased bandwidth, lower unified memory requirements, etc. Tools for distillation 19-3 can provide for the training of lighter-weight models based on the knowledge encoded in development model 16. For instance, development model 16 can be a highly performant, large machine-learned model optimized using model development platform 12. To obtain a lightweight model for running in resource-constrained environments, a smaller model can be a “student model” that learns to imitate development model 16 as a “teacher model.” In this manner, for instance, the investment in learning the parameters and configurations of development model 16 can be efficiently transferred to a smaller model for more efficient inference.

[0185] Workbench 15 can implement one, multiple, or none of the toolkits implemented in model development platform 12. Workbench 15 can output an output model 20 based on development model 16. Output model 20 can be a deployment version of development model 16. Output model 20 can be a development or training checkpoint of development model 16. Output model 20 can be a distilled, compressed, or otherwise optimized version of development model 16.

[0186] Figure 14 is a block diagram of an example training flow for training a machine-learned development model 16. One or more portion(s) of the example training flow can be implemented by a computing system that includes one or more computing devices such as, for example, computing systems described with reference to the other figures. Each respective portion of the example training flow can be performed by any (or any combination) of one or more computing devices. Moreover, one or more portion(s) of the example training flow can be implemented on the hardware components of the device(s) described herein, for example, to train one or more systems or models. FIG. 14 depicts elements performed in a particular order for purposes of illustration and discussion. Those of ordinary skill in the art, using the disclosures provided herein, will understand that the elements of any of the methods discussed herein can be adapted, rearranged, expanded, omitted, combined, or modified in various ways without deviating from the scope of the present disclosure. FIG. 14 is described with reference to elements/terms described with respect to other systems and figures for exemplary illustrated purposes and is not meant to be limiting. One or more portions of the example training flow can be performed additionally, or alternatively, by other systems.

[0187] Initially, development model 16 can persist in an initial state as an initialized model 21. Development model 16 can be initialized with weight values. Initial weight values can be random or based on an initialization schema. Initial weight values can be based on prior pre-training for the same or for a different model.

[0188] Initialized model 21 can undergo pre-training in a pre-training stage 22. Pretraining stage 22 can be implemented using one or more pre-training pipelines 17-2 over data from dataset(s) 17-1. Pre-training can be omitted, for example, if initialized model 21 is already pre-trained (e.g., development model 16 contains, is, or is based on a pre-trained foundational model or an expert model).

[0189] Pre-trained model 23 can then be a new version of development model 16, which can persist as development model 16 or as a new development model. Pre-trained model 23 can be the initial state if development model 16 was already pre-trained. Pre-trained model 23 can undergo fine-tuning in a fine-tuning stage 24. Fine-tuning stage 24 can be implemented using one or more fine-tuning pipelines 17-3 over data from dataset(s) 17-1. Fine-tuning can be omitted, for example, if a pre-trained model has satisfactory performance, if the model was already fine-tuned, or if other tuning approaches are preferred.

[0190] Fine-tuned model 29 can then be a new version of development model 16, which can persist as development model 16 or as a new development model. Fine-tuned model 29 can be the initial state if development model 16 was already fine-tuned. Fine-tuned model 29 can undergo refinement with user feedback 26. For instance, refinement with user feedback 26 can include reinforcement learning, optionally based on human feedback from human users of fine-tuned model 25. As reinforcement learning can be a form of fine-tuning, it is to be understood that fine-tuning stage 24 can subsume the stage for refining with user feedback 26. Refinement with user feedback 26 can produce a refined model 27. Refined model 27 can be output to downstream system(s) 28 for deployment or further development. [0191] In some implementations, computational optimization operations can be applied before, during, or after each stage. For instance, initialized model 21 can undergo computational optimization 29-1 (e.g., using computational optimization toolkit 19) before pre-training stage 22. Pre-trained model 23 can undergo computational optimization 29-2 (e.g., using computational optimization toolkit 19) before fine-tuning stage 24. Fine-tuned model 25 can undergo computational optimization 29-3 (e.g., using computational optimization toolkit 19) before refinement with user feedback 26. Refined model 27 can undergo computational optimization 29-4 (e.g., using computational optimization toolkit 19) before output to downstream system(s) 28. Computational optimization(s) 29-1, . . . , 29-4 can all be the same, all be different, or include at least some different optimization techniques.

[0192] Figure 15 is a block diagram of an inference system for operating one or more machine-learned model(s) 1 to perform inference (e.g., for training, for deployment, etc.). A model host 31 can receive machine-learned model(s) 1. Model host 31 can host one or more model instance(s) 31-1, which can be one or multiple instances of one or multiple models. Model host 31 can host model instance(s) 31-1 using available compute resources 31-2 associated with model host 31.

[0193] Model host 31 can perform inference on behalf of one or more client(s) 32.

Client(s) 32 can transmit an input request 33 to model host 31. Using input request 33, model host 31 can obtain input(s) 2 for input to machine-learned model(s) 1. Machine-learned model(s) 1 can process input(s) 2 to generate output(s) 3. Using output(s) 3, model host 31 can return an output payload 34 for responding to input request 33 from client(s) 32. Output payload 34 can include or be based on output(s) 3.

[0194] Model host 31 can leverage various other resources and tools to augment the inference task. For instance, model host 31 can communicate with tool interfaces 35 to facilitate tool use by model instance(s) 31-1. Tool interfaces 35 can include local or remote APIs. Tool interfaces 35 can include integrated scripts or other software functionality. Model host 31 can engage online learning interface(s) 36 to facilitate ongoing improvements to machine-learned model(s) 1. For instance, online learning interface(s) 36 can be used within reinforcement learning loops to retrieve user feedback on inferences served by model host 31. Model host 31 can access runtime data source(s) 37 for augmenting input(s) 2 with additional contextual information. For instance, runtime data source(s) 37 can include a knowledge graph 37-1 that facilitates structured information retrieval for information associated with input request(s) 33 (e.g., a search engine service). Runtime data source(s) 37 can include public or private, external or local database(s) 37-2 that can store information associated with input request(s) 33 for augmenting input(s) 2. Runtime data source(s) 37 can include account data 37-3 which can be retrieved in association with a user account corresponding to a client 32 for customizing the behavior of model host 31 accordingly.

[0195] Model host 31 can be implemented by one or multiple computing devices or systems. Client(s) 2 can be implemented by one or multiple computing devices or systems, which can include computing devices or systems shared with model host 31. [0196] For example, model host 31 can operate on a server system that provides a machine-learning service to client device(s) that operate client(s) 32 (e.g., over a local or wide-area network). Client device(s) can be end-user devices used by individuals. Client device(s) can be server systems that operate client(s) 32 to provide various functionality as a service to downstream end-user devices.

[0197] In some implementations, model host 31 can operate on a same device or system as client(s) 32. Model host 31 can be a machine-learning service that runs on-device to provide machine-learning functionality to one or multiple applications operating on a client device, which can include an application implementing client(s) 32. Model host 31 can be a part of a same application as client(s) 32. For instance, model host 31 can be a subroutine or method implemented by one part of an application, and client(s) 32 can be another subroutine or method that engages model host 31 to perform inference functions within the application. It is to be understood that model host 31 and client(s) 32 can have various different configurations.

[0198] Model instance(s) 31-1 can include one or more machine-learned models that are available for performing inference. Model instance(s) 31-1 can include weights or other model components that are stored on/in persistent storage, temporarily cached, or loaded into high-speed memory. Model instance(s) 31-1 can include multiple instance(s) of the same model (e.g., for parallel execution of more requests on the same model). Model instance(s) 31-1 can include instance(s) of different model(s). Model instance(s) 31-1 can include cached intermediate states of active or inactive model(s) used to accelerate inference of those models. For instance, an inference session with a particular model may generate significant amounts of computational results that can be re-used for future inference runs (e.g., using a KV cache for transformer-based models). These computational results can be saved in association with that inference session so that session can be executed more efficiently when resumed.

[0199] Compute resource(s) 31-2 can include one or more processors (central processing units, graphical processing units, tensor processing units, machine-learning accelerators, etc.) connected to one or more memory devices. Compute resource(s) 31-2 can include a dynamic pool of available resources shared with other processes. Compute resource(s) 31-2 can include memory devices large enough to fit an entire model instance in a single memory instance. Compute resource(s) 31-2 can also share model instance(s) across multiple memory devices (e.g., using data parallelization or tensor parallelization, etc.). This can be done to increase parallelization or to execute a large model using multiple memory devices which individually might not be able to fit the entire model into memory.

[0200] Input request 33 can include data for input(s) 2. Model host 31 can process input request 33 to obtain input(s) 2. Input(s) 2 can be obtained directly from input request 33 or can be retrieved using input request 33. Input request 33 can be submitted to model host 31 via an API.

[0201] Model host 31 can perform inference over batches of input requests 33 in parallel. For instance, a model instance 31-1 can be configured with an input structure that has a batch dimension. Separate input(s) 2 can be distributed across the batch dimension (e.g., rows of an array). The separate input(s) 2 can include completely different contexts. The separate input(s) 2 can be multiple inference steps of the same task. The separate input(s) 2 can be staggered in an input structure, such that any given inference cycle can be operating on different portions of the respective input(s) 2. In this manner, for instance, model host 31 can perform inference on the batch in parallel, such that output(s) 3 can also contain the batch dimension and return the inference results for the batched input(s) 2 in parallel. In this manner, for instance, batches of input request(s) 33 can be processed in parallel for higher throughput of output payload(s) 34.

[0202] Output payload 34 can include or be based on output(s) 3 from machine- learned model(s) 1. Model host 31 can process output(s) 3 to obtain output payload 34. This can include chaining multiple rounds of inference (e.g., iteratively, recursively, across the same model(s) or different model(s)) to arrive at a final output for a task to be returned in output payload 34. Output payload 34 can be transmitted to client(s) 32 via an API.

[0203] Online learning interface(s) 36 can facilitate reinforcement learning of machine-learned model(s) 1. Online learning interface(s) 36 can facilitate reinforcement learning with human feedback (RLHF). Online learning interface(s) 36 can facilitate federated learning of machine-learned model(s) 1.

[0204] Model host 31 can execute machine-learned model(s) 1 to perform inference for various tasks using various types of data. For example, various different input(s) 2 and output(s) 3 can be used for various different tasks. In some implementations, input(s) 2 can be or otherwise represent image data. Machine-learned model(s) 1 can process the image data to generate an output. As an example, machine-learned model(s) 1 can process the image data to generate an image recognition output (e.g., a recognition of the image data, a latent embedding of the image data, an encoded representation of the image data, a hash of the image data, etc.). As another example, machine-learned model(s) 1 can process the image data to generate an image segmentation output. As another example, machine-learned model(s) 1 can process the image data to generate an image classification output. As another example, machine-learned model(s) 1 can process the image data to generate an image data modification output (e.g., an alteration of the image data, etc.). As another example, machine- learned model(s) 1 can process the image data to generate an encoded image data output (e.g., an encoded and/or compressed representation of the image data, etc.). As another example, machine-learned model(s) 1 can process the image data to generate an upscaled image data output. As another example, machine-learned model(s) 1 can process the image data to generate a prediction output.

[0205] In some implementations, the task is a computer vision task. In some cases, input(s) 2 includes pixel data for one or more images and the task is an image processing task. For example, the image processing task can be image classification, where the output is a set of scores, each score corresponding to a different object class and representing the likelihood that the one or more images depict an object belonging to the object class. The image processing task may be object detection, where the image processing output identifies one or more regions in the one or more images and, for each region, a likelihood that region depicts an object of interest. As another example, the image processing task can be image segmentation, where the image processing output defines, for each pixel in the one or more images, a respective likelihood for each category in a predetermined set of categories. For example, the set of categories can be foreground and background. As another example, the set of categories can be object classes. As another example, the image processing task can be depth estimation, where the image processing output defines, for each pixel in the one or more images, a respective depth value. As another example, the image processing task can be motion estimation, where the network input includes multiple images, and the image processing output defines, for each pixel of one of the input images, a motion of the scene depicted at the pixel between the images in the network input.

[0206] In some implementations, input(s) 2 can be or otherwise represent natural language data. Machine-learned model(s) 1 can process the natural language data to generate an output. As an example, machine-learned model(s) 1 can process the natural language data to generate a language encoding output. As another example, machine-learned model(s) 1 can process the natural language data to generate a latent text embedding output. As another example, machine-learned model(s) 1 can process the natural language data to generate a translation output. As another example, machine-learned model(s) 1 can process the natural language data to generate a classification output. As another example, machine-learned model(s) 1 can process the natural language data to generate a textual segmentation output. As another example, machine-learned model(s) 1 can process the natural language data to generate a semantic intent output. As another example, machine-learned model(s) 1 can process the natural language data to generate an upscaled text or natural language output (e.g., text or natural language data that is higher quality than the input text or natural language, etc.). As another example, machine-learned model(s) 1 can process the natural language data to generate a prediction output (e.g., one or more predicted next portions of natural language content).

[0207] In some implementations, input(s) 2 can be or otherwise represent speech data (e.g., data describing spoken natural language, such as audio data, textual data, etc.). Machine-learned model(s) 1 can process the speech data to generate an output. As an example, machine-learned model(s) 1 can process the speech data to generate a speech recognition output. As another example, machine-learned model(s) 1 can process the speech data to generate a speech translation output. As another example, machine-learned model(s) 1 can process the speech data to generate a latent embedding output. As another example, machine-learned model(s) 1 can process the speech data to generate an encoded speech output (e.g., an encoded and/or compressed representation of the speech data, etc.). As another example, machine-learned model(s) 1 can process the speech data to generate an upscaled speech output (e.g., speech data that is higher quality than the input speech data, etc.). As another example, machine-learned model(s) 1 can process the speech data to generate a textual representation output (e.g., a textual representation of the input speech data, etc.). As another example, machine-learned model(s) 1 can process the speech data to generate a prediction output.

[0208] In some implementations, input(s) 2 can be or otherwise represent latent encoding data (e.g., a latent space representation of an input, etc.). Machine-learned model(s) 1 can process the latent encoding data to generate an output. As an example, machine- learned model(s) 1 can process the latent encoding data to generate a recognition output. As another example, machine-learned model(s) 1 can process the latent encoding data to generate a reconstruction output. As another example, machine-learned model(s) 1 can process the latent encoding data to generate a search output. As another example, machine- learned model(s) 1 can process the latent encoding data to generate a reclustering output. As another example, machine-learned model(s) 1 can process the latent encoding data to generate a prediction output. [0209] In some implementations, input(s) 2 can be or otherwise represent statistical data. Statistical data can be, represent, or otherwise include data computed and/or calculated from some other data source. Machine-learned model(s) 1 can process the statistical data to generate an output. As an example, machine-learned model(s) 1 can process the statistical data to generate a recognition output. As another example, machine-learned model(s) 1 can process the statistical data to generate a prediction output. As another example, machine- learned model(s) 1 can process the statistical data to generate a classification output. As another example, machine-learned model(s) 1 can process the statistical data to generate a segmentation output. As another example, machine-learned model(s) 1 can process the statistical data to generate a visualization output. As another example, machine-learned model(s) 1 can process the statistical data to generate a diagnostic output.

[0210] In some implementations, input(s) 2 can be or otherwise represent sensor data. Machine-learned model(s) 1 can process the sensor data to generate an output. As an example, machine-learned model(s) 1 can process the sensor data to generate a recognition output. As another example, machine-learned model(s) 1 can process the sensor data to generate a prediction output. As another example, machine-learned model(s) 1 can process the sensor data to generate a classification output. As another example, machine-learned model(s) 1 can process the sensor data to generate a segmentation output. As another example, machine-learned model(s) 1 can process the sensor data to generate a visualization output. As another example, machine-learned model(s) 1 can process the sensor data to generate a diagnostic output. As another example, machine-learned model(s) 1 can process the sensor data to generate a detection output.

[0211] In some implementations, machine-learned model(s) 1 can be configured to perform a task that includes encoding input data for reliable and/or efficient transmission or storage (and/or corresponding decoding). For example, the task may be an audio compression task. The input may include audio data and the output may include compressed audio data. In another example, the input includes visual data (e.g. one or more images or videos), the output includes compressed visual data, and the task is a visual data compression task. In another example, the task may include generating an embedding for input data (e.g. input audio or visual data). In some cases, the input includes audio data representing a spoken utterance and the task is a speech recognition task. The output may include a text output which is mapped to the spoken utterance. In some cases, the task includes encrypting or decrypting input data. In some cases, the task includes a microprocessor performance task, such as branch prediction or memory address translation. [0212] In some implementations, the task is a generative task, and machine-learned model(s) 1 can be configured to output content generated in view of input(s) 2. For instance, input(s) 2 can be or otherwise represent data of one or more modalities that encodes context for generating additional content.

[0213] In some implementations, the task can be a text completion task. Machine- learned model(s) 1 can be configured to process input(s) 2 that represent textual data and to generate output(s) 3 that represent additional textual data that completes a textual sequence that includes input(s) 2. For instance, machine-learned model(s) 1 can be configured to generate output(s) 3 to complete a sentence, paragraph, or portion of text that follows from a portion of text represented by input(s) 2.

[0214] In some implementations, the task can be an instruction following task. Machine-learned model(s) 1 can be configured to process input(s) 2 that represent instructions to perform a function and to generate output(s) 3 that advance a goal of satisfying the instruction function (e.g., at least a step of a multi-step procedure to perform the function). Output(s) 3 can represent data of the same or of a different modality as input(s) 2. For instance, input(s) 2 can represent textual data (e.g., natural language instructions for a task to be performed) and machine-learned model(s) 1 can process input(s) 2 to generate output(s) 3 that represent textual data responsive to the instructions (e.g., natural language responses, programming language responses, machine language responses, etc.). Input(s) 2 can represent image data (e.g., image-based instructions for a task to be performed, optionally accompanied by textual instructions) and machine-learned model(s) 1 can process input(s) 2 to generate output(s) 3 that represent textual data responsive to the instructions (e.g., natural language responses, programming language responses, machine language responses, etc.). One or more output(s) 3 can be iteratively or recursively generated to sequentially process and accomplish steps toward accomplishing the requested functionality. For instance, an initial output can be executed by an external system or be processed by machine-learned model(s) 1 to complete an initial step of performing a function. Multiple steps can be performed, with a final output being obtained that is responsive to the initial instructions.

[0215] In some implementations, the task can be a question answering task. Machine- learned model(s) 1 can be configured to process input(s) 2 that represent a question to answer and to generate output(s) 3 that advance a goal of returning an answer to the question (e.g., at least a step of a multi-step procedure to perform the function). Output(s) 3 can represent data of the same or of a different modality as input(s) 2. For instance, input(s) 2 can represent textual data (e.g., natural language instructions for a task to be performed) and machine- learned model(s) 1 can process input(s) 2 to generate output(s) 3 that represent textual data responsive to the question (e.g., natural language responses, programming language responses, machine language responses, etc.). Input(s) 2 can represent image data (e.g., image-based instructions for a task to be performed, optionally accompanied by textual instructions) and machine-learned model(s) 1 can process input(s) 2 to generate output(s) 3 that represent textual data responsive to the question (e.g., natural language responses, programming language responses, machine language responses, etc.). One or more output(s) 3 can be iteratively or recursively generated to sequentially process and accomplish steps toward answering the question. For instance, an initial output can be executed by an external system or be processed by machine-learned model(s) 1 to complete an initial step of obtaining an answer to the question (e.g., querying a database, performing a computation, executing a script, etc.). Multiple steps can be performed, with a final output being obtained that is responsive to the question.

[0216] In some implementations, the task can be an image generation task. Machine- learned model(s) 1 can be configured to process input(s) 2 that represent context regarding a desired portion of image content. The context can include text data, image data, audio data, etc. Machine-learned model(s) 1 can be configured to generate output(s) 3 that represent image data that depicts imagery related to the context. For instance, machine-learned model(s) 1 can be configured to generate pixel data of an image. Values for channel(s) associated with the pixels in the pixel data can be selected based on the context (e.g., based on a probability determined based on the context).

[0217] In some implementations, the task can be an audio generation task. Machine- learned model(s) 1 can be configured to process input(s) 2 that represent context regarding a desired portion of audio content. The context can include text data, image data, audio data, etc. Machine-learned model(s) 1 can be configured to generate output(s) 3 that represent audio data related to the context. For instance, machine-learned model(s) 1 can be configured to generate waveform data in the form of an image (e.g., a spectrogram). Values for channel(s) associated with pixels of the image can be selected based on the context. Machine- learned model(s) 1 can be configured to generate waveform data in the form of a sequence of discrete samples of a continuous waveform. Values of the sequence can be selected based on the context (e.g., based on a probability determined based on the context).

[0218] In some implementations, the task can be a data generation task. Machine- learned model(s) 1 can be configured to process input(s) 2 that represent context regarding a desired portion of data (e.g., data from various data domains, such as sensor data, image data, multimodal data, statistical data, etc.). The desired data can be, for instance, synthetic data for training other machine-learned models. The context can include arbitrary data type(s). Machine-learned model(s) 1 can be configured to generate output(s) 3 that represent data that aligns with the desired data. For instance, machine-learned model(s) 1 can be configured to generate data values for populating a dataset. Values for the data object(s) can be selected based on the context (e.g., based on a probability determined based on the context).

[0219] Figure 16 is a block diagram of an example networked computing system that can perform aspects of example implementations of the present disclosure. The system can include a number of computing devices and systems that are communicatively coupled over a network 49. An example computing device 50 is described to provide an example of a computing device that can perform any aspect of the present disclosure (e.g., implementing model host 31, client(s) 32, or both). An example server computing system 60 is described as an example of a server computing system that can perform any aspect of the present disclosure (e.g., implementing model host 31, client(s) 32, or both). Computing device 50 and server computing system(s) 60 can cooperatively interact (e.g., over network 49) to perform any aspect of the present disclosure (e.g., implementing model host 31, client(s) 32, or both). Model development platform system 70 is an example system that can host or serve model development platform(s) 12 for development of machine-learned models. Third-party system(s) 80 are example system(s) with which any of computing device 50, server computing system(s) 60, or model development platform system(s) 70 can interact in the performance of various aspects of the present disclosure (e.g., engaging third-party tools, accessing third-party databases or other resources, etc.).

[0220] Network 49 can be any type of communications network, such as a local area network (e.g., intranet), wide area network (e.g., Internet), or some combination thereof and can include any number of wired or wireless links. In general, communication over network 49 can be carried via any type of wired or wireless connection, using a wide variety of communication protocols (e.g., TCP/IP, HTTP, SMTP, FTP), encodings or formats (e.g., HTML, XML), or protection schemes (e.g., VPN, secure HTTP, SSL). Network 49 can also be implemented via a system bus. For instance, one or more devices or systems of Figure 16 can be co-located with, contained by, or otherwise integrated into one or more other devices or systems.

[0221] Computing device 50 can be any type of computing device, such as, for example, a personal computing device (e.g., laptop or desktop), a mobile computing device (e.g., smartphone or tablet), a gaming console or controller, a wearable computing device, an embedded computing device, a server computing device, a virtual machine operating on a host device, or any other type of computing device. Computing device 50 can be a client computing device. Computing device 50 can be an end-user computing device. Computing device 50 can be a computing device of a service provided that provides a service to an end user (who may use another computing device to interact with computing device 50).

[0222] Computing device 50 can include one or more processors 51 and a memory 52. Processor(s) 51 can be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, an FPGA, a controller, a microcontroller, etc.) and can be one processor or a plurality of processors that are operatively connected. Memory 52 can include one or more non-transitory computer-readable storage media, such as HBM, RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. Memory 52 can store data 53 and instructions 54 which can be executed by processor(s) 51 to cause computing device 50 to perform operations. The operations can implement any one or multiple features described herein. The operations can implement example methods and techniques described herein.

[0223] Computing device 50 can also include one or more input components that receive user input. For example, a user input component can be a touch-sensitive component (e.g., a touch-sensitive display screen or a touch pad) that is sensitive to the touch of a user input object (e.g., a finger or a stylus). The touch-sensitive component can serve to implement a virtual keyboard. Other example user input components include a microphone, camera, LIDAR, a physical keyboard or other buttons, or other means by which a user can provide user input.

[0224] Computing device 50 can store or include one or more machine-learned models 55. Machine-learned models 55 can include one or more machine-learned model(s) 1, such as a sequence processing model 4. Machine-learned models 55 can include one or multiple model instance(s) 31-1. Machine-learned model(s) 55 can be received from server computing system(s) 60, model development platform system 70, third party system(s) 80 (e.g., an application distribution platform), or developed locally on computing device 50. Machine-learned model(s) 55 can be loaded into memory 52 and used or otherwise implemented by processor(s) 51. Computing device 50 can implement multiple parallel instances of machine-learned model(s) 55.

[0225] Server computing system(s) 60 can include one or more processors 61 and a memory 62. Processor(s) 61 can be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, an FPGA, a controller, a microcontroller, etc.) and can be one processor or a plurality of processors that are operatively connected. Memory 62 can include one or more non-transitory computer-readable storage media, such as HBM, RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. Memory 62 can store data 63 and instructions 64 which can be executed by processor(s) 61 to cause server computing system(s) 60 to perform operations. The operations can implement any one or multiple features described herein. The operations can implement example methods and techniques described herein.

[0226] In some implementations, server computing system 60 includes or is otherwise implemented by one or multiple server computing devices. In instances in which server computing system 60 includes multiple server computing devices, such server computing devices can operate according to sequential computing architectures, parallel computing architectures, or some combination thereof.

[0227] Server computing system 60 can store or otherwise include one or more machine-learned models 65. Machine-learned model(s) 65 can be the same as or different from machine-learned model(s) 55. Machine-learned models 65 can include one or more machine-learned model(s) 1, such as a sequence processing model 4. Machine-learned models 65 can include one or multiple model instance(s) 31-1. Machine-learned model(s) 65 can be received from computing device 50, model development platform system 70, third party system(s) 80, or developed locally on server computing system(s) 60. Machine-learned model(s) 65 can be loaded into memory 62 and used or otherwise implemented by processor(s) 61. Server computing system(s) 60 can implement multiple parallel instances of machine-learned model (s) 65.

[0228] In an example configuration, machine-learned models 65 can be included in or otherwise stored and implemented by server computing system 60 to establish a client-server relationship with computing device 50 for serving model inferences. For instance, server computing system(s) 60 can implement model host 31 on behalf of client(s) 32 on computing device 50. For instance, machine-learned models 65 can be implemented by server computing system 60 as a portion of a web service (e.g., remote machine-learned model hosting service, such as an online interface for performing machine-learned model operations over a network on server computing system(s) 60). For instance, server computing system(s) 60 can communicate with computing device 50 over a local intranet or internet connection. For instance, computing device 50 can be a workstation or endpoint in communication with server computing system(s) 60, with implementation of machine-learned models 65 being managed by server computing system(s) 60 to remotely perform inference (e.g., for runtime or training operations), with output(s) returned (e.g., cast, streamed, etc.) to computing device 50. Machine-learned models 65 can work cooperatively or interoperatively with machine- learned models 55 on computing device 50 to perform various tasks.

[0229] Model development platform system(s) 70 can include one or more processors 71 and a memory 72. Processor(s) 71 can be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, an FPGA, a controller, a microcontroller, etc.) and can be one processor or a plurality of processors that are operatively connected. Memory 72 can include one or more non-transitory computer-readable storage media, such as HBM, RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. Memory 72 can store data 73 and instructions 74 which can be executed by processor(s) 71 to cause model development platform system(s) 70 to perform operations. The operations can implement any one or multiple features described herein. The operations can implement example methods and techniques described herein. Example operations include the functionality described herein with respect to model development platform 12. This and other functionality can be implemented by developer tool(s) 75.

[0230] Third-party system(s) 80 can include one or more processors 81 and a memory 82. Processor(s) 81 can be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, an FPGA, a controller, a microcontroller, etc.) and can be one processor or a plurality of processors that are operatively connected. Memory 82 can include one or more non-transitory computer-readable storage media, such as HBM, RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. Memory 82 can store data 83 and instructions 84 which can be executed by processor(s) 81 to cause third-party system(s) 80 to perform operations. The operations can implement any one or multiple features described herein. The operations can implement example methods and techniques described herein. Example operations include the functionality described herein with respect to tools and other external resources called when training or performing inference with machine-learned model(s) 1, 4, 16, 20, 55, 65, etc. (e.g., third-party resource(s) 85).

[0231] Figure 16 illustrates one example arrangement of computing systems that can be used to implement the present disclosure. Other computing system configurations can be used as well. For example, in some implementations, one or both of computing system 50 or server computing system(s) 60 can implement all or a portion of the operations of model development platform system 70. For example, computing system 50 or server computing system(s) 60 can implement developer tool(s) 75 (or extensions thereof) to develop, update/train, or refine machine-learned models 1, 4, 16, 20, 55, 65, etc. using one or more techniques described herein with respect to model alignment toolkit 17. In this manner, for instance, computing system 50 or server computing system(s) 60 can develop, update/train, or refine machine-learned models based on local datasets (e.g., for model personalization/customization, as permitted by user data preference selections).

[0232] Figure 17 is a block diagram of an example computing device 98 that performs according to example embodiments of the present disclosure. Computing device 98 can be a user computing device or a server computing device (e.g., computing device 50, server computing system(s) 60, etc.). Computing device 98 can implement model host 31. For instance, computing device 98 can include a number of applications (e.g., applications 1 through N). Each application can contain its own machine learning library and machine- learned model(s). For example, each application can include a machine-learned model. Example applications include a text messaging application, an email application, a dictation application, a virtual keyboard application, a browser application, etc. As illustrated in Figure 17, each application can communicate with a number of other components of the computing device, such as, for example, one or more sensors, a context manager, a device state component, or additional components. In some implementations, each application can communicate with each device component using an API (e.g., a public API). In some implementations, the API used by each application is specific to that application.

[0233] Figure 18 is a block diagram of an example computing device 99 that performs according to example embodiments of the present disclosure. Computing device 99 can be the same as or different from computing device 98. Computing device 99 can be a user computing device or a server computing device (e.g., computing device 50, server computing system(s) 60, etc.). Computing device 98 can implement model host 31. For instance, computing device 99 can include a number of applications (e.g., applications 1 through N). Each application can be in communication with a central intelligence layer. Example applications include a text messaging application, an email application, a dictation application, a virtual keyboard application, a browser application, etc. In some implementations, each application can communicate with the central intelligence layer (and model(s) stored therein) using an API (e.g., a common API across all applications).

[0234] The central intelligence layer can include a number of machine-learned models. For example, as illustrated in Figure 18, a respective machine-learned model can be provided for each application and managed by the central intelligence layer. In other implementations, two or more applications can share a single machine-learned model. For example, in some implementations, the central intelligence layer can provide a single model for all of the applications. In some implementations, the central intelligence layer is included within or otherwise implemented by an operating system of computing device 99.

[0235] The central intelligence layer can communicate with a central device data layer. The central device data layer can be a centralized repository of data for computing device 99. As illustrated in Figure 18, the central device data layer can communicate with a number of other components of the computing device, such as, for example, one or more sensors, a context manager, a device state component, or additional components. In some implementations, the central device data layer can communicate with each device component using an API (e.g., a private API).

[0236] Figure 19A depicts a block diagram of an example computing system 1000 that performs document-to-dialogue audio conversion according to example embodiments of the present disclosure. The system 1000 includes a user computing system 1002, a server computing system 1030, and/or a third computing system 1050 that are communicatively coupled over a network 1080.

[0237] The user computing system 1002 can include any type of computing device, such as, for example, a personal computing device (e.g., laptop or desktop), a mobile computing device (e.g., smartphone or tablet), a gaming console or controller, a wearable computing device, an embedded computing device, or any other type of computing device. [0238] The user computing system 1002 includes one or more processors 1012 and a memory 1014. The one or more processors 1012 can be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, a FPGA, a controller, a microcontroller, etc.) and can be one processor or a plurality of processors that are operatively connected. The memory 1014 can include one or more non-transitory computer-readable storage mediums, such as RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. The memory 1014 can store data 1016 and instructions 1018 which are executed by the processor 1012 to cause the user computing system 1002 to perform operations.

[0239] In some implementations, the user computing system 1002 can store or include one or more machine-learned models 1020. For example, the machine-learned models 1020 can be or can otherwise include various machine-learned models such as neural networks (e.g., deep neural networks) or other types of machine-learned models, including non-linear models and/or linear models. Neural networks can include feed-forward neural networks, recurrent neural networks (e.g., long short-term memory recurrent neural networks), convolutional neural networks or other forms of neural networks.

[0240] In some implementations, the one or more machine-learned models 120 can be received from the server computing system 1030 over network 1080, stored in the user computing device memory 1014, and then used or otherwise implemented by the one or more processors 1012. In some implementations, the user computing system 1002 can implement multiple parallel instances of a single machine-learned model 1020 (e.g., to perform parallel machine-learned model processing across multiple instances of input data and/or detected features).

[0241] More particularly, the one or more machine-learned models 1020 may include one or more detection models, one or more classification models, one or more segmentation models, one or more augmentation models, one or more generative models, one or more natural language processing models, one or more optical character recognition models, and/or one or more other machine-learned models. The one or more machine-learned models 1020 can include one or more transformer models. The one or more machine-learned models 1020 may include one or more neural radiance field models, one or more diffusion models, and/or one or more autoregressive language models.

[0242] The one or more machine-learned models 1020 may be utilized to detect one or more object features. The detected object features may be classified and/or embedded. The classification and/or the embedding may then be utilized to perform a search to determine one or more search results. Alternatively and/or additionally, the one or more detected features may be utilized to determine an indicator (e.g., a user interface element that indicates a detected feature) is to be provided to indicate a feature has been detected. The user may then select the indicator to cause a feature classification, embedding, and/or search to be performed. In some implementations, the classification, the embedding, and/or the searching can be performed before the indicator is selected.

[0243] In some implementations, the one or more machine-learned models 1020 can process image data, text data, audio data, and/or latent encoding data to generate output data that can include image data, text data, audio data, and/or latent encoding data. The one or more machine-learned models 1020 may perform optical character recognition, natural language processing, image classification, object classification, text classification, audio classification, context determination, action prediction, image correction, image augmentation, text augmentation, sentiment analysis, object detection, error detection, inpainting, video stabilization, audio correction, audio augmentation, and/or data segmentation (e.g., mask based segmentation).

[0244] Additionally or alternatively, one or more machine-learned models 1040 can be included in or otherwise stored and implemented by the server computing system 1030 that communicates with the user computing system 1002 according to a client-server relationship. For example, the machine-learned models 1040 can be implemented by the server computing system 1030 as a portion of a web service (e.g., a viewfinder service, a visual search service, an image processing service, an ambient computing service, and/or an overlay application service). Thus, one or more models 1020 can be stored and implemented at the user computing system 1002 and/or one or more models 1040 can be stored and implemented at the server computing system 1030.

[0245] The user computing system 1002 can also include one or more user input component 1022 that receives user input. For example, the user input component 1022 can be a touch-sensitive component (e.g., a touch-sensitive display screen or a touch pad) that is sensitive to the touch of a user input object (e.g., a finger or a stylus). The touch-sensitive component can serve to implement a virtual keyboard. Other example user input components include a microphone, a traditional keyboard, or other means by which a user can provide user input.

[0246] In some implementations, the user computing system 1002 can store and/or provide one or more user interfaces 1024, which may be associated with one or more applications. The one or more user interfaces 1024 can be configured to receive inputs and/or provide data for display (e.g., image data, text data, audio data, one or more user interface elements, an augmented-reality experience, a virtual reality experience, and/or other data for display. The user interfaces 1024 may be associated with one or more other computing systems (e.g., server computing system 1030 and/or third party computing system 1050). The user interfaces 1024 can include a viewfinder interface, a search interface, a generative model interface, a social media interface, and/or a media content gallery interface.

[0247] The user computing system 1002 may include and/or receive data from one or more sensors 1026. The one or more sensors 1026 may be housed in a housing component that houses the one or more processors 1012, the memory 1014, and/or one or more hardware components, which may store, and/or cause to perform, one or more software packets. The one or more sensors 1026 can include one or more image sensors (e.g., a camera), one or more lidar sensors, one or more audio sensors (e.g., a microphone), one or more inertial sensors (e.g., inertial measurement unit), one or more biological sensors (e.g., a heart rate sensor, a pulse sensor, a retinal sensor, and/or a fingerprint sensor), one or more infrared sensors, one or more location sensors (e.g., GPS), one or more touch sensors (e.g., a conductive touch sensor and/or a mechanical touch sensor), and/or one or more other sensors. The one or more sensors can be utilized to obtain data associated with a user’s environment (e.g., an image of a user’s environment, a recording of the environment, and/or the location of the user).

[0248] The user computing system 1002 may include, and/or be part of, a user computing device 1004. The user computing device 1004 may include a mobile computing device (e.g., a smartphone or tablet), a desktop computer, a laptop computer, a smart wearable, and/or a smart appliance. Additionally and/or alternatively, the user computing system may obtain from, and/or generate data with, the one or more one or more user computing devices 1004. For example, a camera of a smartphone may be utilized to capture image data descriptive of the environment, and/or an overlay application of the user computing device 1004 can be utilized to track and/or process the data being provided to the user. Similarly, one or more sensors associated with a smart wearable may be utilized to obtain data about a user and/or about a user’s environment (e.g., image data can be obtained with a camera housed in a user’s smart glasses). Additionally and/or alternatively, the data may be obtained and uploaded from other user devices that may be specialized for data obtainment or generation.

[0249] The server computing system 1030 includes one or more processors 1032 and a memory 1034. The one or more processors 1032 can be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, a FPGA, a controller, a microcontroller, etc.) and can be one processor or a plurality of processors that are operatively connected. The memory 1034 can include one or more non-transitory computer-readable storage mediums, such as RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. The memory 1034 can store data 1036 and instructions 1038 which are executed by the processor 1032 to cause the server computing system 1030 to perform operations.

[0250] In some implementations, the server computing system 1030 includes or is otherwise implemented by one or more server computing devices. In instances in which the server computing system 1030 includes plural server computing devices, such server computing devices can operate according to sequential computing architectures, parallel computing architectures, or some combination thereof. [0251] As described above, the server computing system 1030 can store or otherwise include one or more machine-learned models 1040. For example, the models 1040 can be or can otherwise include various machine-learned models. Example machine-learned models include neural networks or other multi-layer non-linear models. Example neural networks include feed forward neural networks, deep neural networks, recurrent neural networks, and convolutional neural networks. Example models 1040 are discussed with reference to Figure 19B.

[0252] Additionally and/or alternatively, the server computing system 1030 can include and/or be communicatively connected with a search engine 1042 that may be utilized to crawl one or more databases (and/or resources). The search engine 1042 can process data from the user computing system 1002, the server computing system 1030, and/or the third party computing system 1050 to determine one or more search results associated with the input data. The search engine 1042 may perform term based search, label based search, Boolean based searches, image search, embedding based search (e.g., nearest neighbor search), multimodal search, and/or one or more other search techniques.

[0253] The server computing system 1030 may store and/or provide one or more user interfaces 1044 for obtaining input data and/or providing output data to one or more users. The one or more user interfaces 1044 can include one or more user interface elements, which may include input fields, navigation tools, content chips, selectable tiles, widgets, data display carousels, dynamic animation, informational pop-ups, image augmentations, text-to- speech, speech-to-text, augmented-reality, virtual-reality, feedback loops, and/or other interface elements.

[0254] The user computing system 1002 and/or the server computing system 1030 can train the models 1020 and/or 1040 via interaction with the third party computing system 1050 that is communicatively coupled over the network 1080. The third party computing system 1050 can be separate from the server computing system 1030 or can be a portion of the server computing system 1030. Alternatively and/or additionally, the third party computing system 1050 may be associated with one or more web resources, one or more web platforms, one or more other users, and/or one or more contexts.

[0255] An example machine-learned model can include a generative model (e.g., a large language model, a foundation model, a vision language model, an image generation model, a text-to-image model, an audio generation model, and/or other generative models). [0256] In some implementations, the computing system 1000 may utilize one or more soft prompts for conditioning the one or more machine-learned models (1020 and/or 1040) for downstream tasks. The one or more soft prompts can include a set of tunable parameters that can be trained (or tuned) as the parameters of the one or more machine-learned models (1020 and/or 1040) are fixed. The one or more soft prompts 1024 can be trained for a specific task and/or a specific set of tasks. Alternatively and/or additionally, the one or more soft prompts 1024 may be trained to condition the one or more machine-learned models (1020 and/or 1040) to perform inferences for a particular individual, one or more entities, and/or one or more tasks such that the output is tailored for that particular individual, particular entities, and/or particular task. The one or more soft prompts 1024 can be obtained and processed with one or more inputs by the one or more machine-learned models (1020 and/or 1040).

[0257] The one or more soft prompts can include a set of machine-learned weights. In particular, the one or more soft prompts can include weights that were trained to condition a generative model to generate model-generated content with one or more particular attributes. For example, the one or more soft prompts can be utilized by a user to generate content based on the fine-tuning. The one or more soft prompts can be extended to a plurality of tasks. For example, the computing system 1000 may tune the set of parameters on a plurality of different content attributes and/or types. The one or more soft prompts may include a plurality of learned vector representations that may be model-readable.

[0258] A particular soft prompt can be obtained based on a particular task, individual, content type, etc. The particular soft prompt can include a set of learned parameters. The set of learned parameters can be processed with the generative model to generate the modelgenerated image.

[0259] The user computing system 1002 and/or the server computing system 1030 may store one or more soft prompts associated with the particular user and/or particular task. The soft prompt(s) can include a set of parameters. The user computing system 1002 and/or the server computing system 1030 may leverage the set of parameters of the soft prompt(s) and a generative model to generate a model-generated content item. In some implementations, the model-generated content item can be generated based on the set of parameters associated with the particular individual and/or task.

[0260] The utilization of a soft prompt (i.e., a set of parameters that can be processed with a generative model for downstream task conditioning) can reduce the computational cost for parameter tuning for object-specific content generation by reducing the parameters to be tuned. The set of parameters can be limited and may be adjusted while the parameters of the pre-trained generative model stay fixed. The set of parameters of the soft prompt can be utilized to condition the pre-trained generative model (e.g., the machine-learned image generation model and/or language model) for particular downstream tasks (e.g., response generation and/or image rendering).

[0261] In some implementations, the generative language model and/or one or more soft prompts (e.g., a set of machine-learned parameters that can be processed with the input by the generative language model) can be trained to generate content with particular attributes.

[0262] In some implementations, the server computing system 1030 can include a prompt library. The prompt library can store a plurality of prompt templates (e.g., a plurality of hard prompt templates (e.g., text prompt templates)) and/or a plurality of soft prompts. The plurality of prompt templates can include hard prompt templates (e.g., text string data) that may be combined with the user input to generate a more detailed and complete prompt for the generative model to process. The templates can include text descriptive of the request. The templates may be object-specific, user-specific, and/or content-specific. The plurality of prompt templates may include few-shot examples.

[0263] The prompt library can store a plurality of soft prompts. The plurality of soft prompts may be associated with a plurality of different content attributes and/or a plurality of different individuals. The plurality of soft prompts can include learned parameters and/or learned weights that can be processed with the generative model to condition the generative model to generate content items with particular attributes. The plurality of soft prompts may have been tuned by freezing the parameters of a pre-trained generative model, while the parameters of the soft prompt are learned based on a particular task and/or user. The plurality of soft prompts can include a plurality of different soft prompts associated with a plurality of different users and/or a plurality of different sets of users.

[0264] The third party computing system 1050 can include one or more processors 1052 and a memory 1054. The one or more processors 1052 can be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, a FPGA, a controller, a microcontroller, etc.) and can be one processor or a plurality of processors that are operatively connected. The memory 1054 can include one or more non-transitory computer- readable storage mediums, such as RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. The memory 1054 can store data 1056 and instructions 1058 which are executed by the processor 1052 to cause the third party computing system 1050 to perform operations. In some implementations, the third party computing system 1050 includes or is otherwise implemented by one or more server computing devices.

[0265] The network 1080 can be any type of communications network, such as a local area network (e.g., intranet), wide area network (e.g., Internet), or some combination thereof and can include any number of wired or wireless links. In general, communication over the network 1080 can be carried via any type of wired and/or wireless connection, using a wide variety of communication protocols (e.g., TCP/IP, HTTP, SMTP, FTP), encodings or formats (e.g., HTML, XML), and/or protection schemes (e.g., VPN, secure HTTP, SSL).

[0266] The machine-learned models described in this specification may be used in a variety of tasks, applications, and/or use cases.

[0267] In some implementations, the input to the machine-learned model(s) of the present disclosure can be image data. The machine-learned model(s) can process the image data to generate an output. As an example, the machine-learned model(s) can process the image data to generate an image recognition output (e.g., a recognition of the image data, a latent embedding of the image data, an encoded representation of the image data, a hash of the image data, etc.). As another example, the machine-learned model(s) can process the image data to generate an image segmentation output. As another example, the machine- learned model(s) can process the image data to generate an image classification output. As another example, the machine-learned model(s) can process the image data to generate an image data modification output (e.g., an alteration of the image data, etc.). As another example, the machine-learned model(s) can process the image data to generate an encoded image data output (e.g., an encoded and/or compressed representation of the image data, etc.). As another example, the machine-learned model(s) can process the image data to generate an upscaled image data output. As another example, the machine-learned model(s) can process the image data to generate a prediction output.

[0268] In some implementations, the input to the machine-learned model(s) of the present disclosure can be text or natural language data. The machine-learned model(s) can process the text or natural language data to generate an output. As an example, the machine- learned model(s) can process the natural language data to generate a language encoding output. As another example, the machine-learned model(s) can process the text or natural language data to generate a latent text embedding output. As another example, the machine- learned model(s) can process the text or natural language data to generate a translation output. As another example, the machine-learned model(s) can process the text or natural language data to generate a classification output. As another example, the machine-learned model(s) can process the text or natural language data to generate a textual segmentation output. As another example, the machine-learned model(s) can process the text or natural language data to generate a semantic intent output. As another example, the machine-learned model(s) can process the text or natural language data to generate an upscaled text or natural language output (e.g., text or natural language data that is higher quality than the input text or natural language, etc.). As another example, the machine-learned model(s) can process the text or natural language data to generate a prediction output.

[0269] The user computing system may include a number of applications (e.g., applications 1 through N). Each application may include its own respective machine learning library and machine-learned model(s). For example, each application can include a machine- learned model. Example applications include a text messaging application, an email application, a dictation application, a virtual keyboard application, a browser application, etc. [0270] Each application can communicate with a number of other components of the computing device, such as, for example, one or more sensors, a context manager, a device state component, and/or additional components. In some implementations, each application can communicate with each device component using an API (e.g., a public API). In some implementations, the API used by each application is specific to that application.

[0271] The user computing system 1002 can include a number of applications (e.g., applications 1 through N). Each application is in communication with a central intelligence layer. Example applications include a text messaging application, an email application, a dictation application, a virtual keyboard application, a browser application, etc. In some implementations, each application can communicate with the central intelligence layer (and model(s) stored therein) using an API (e.g., a common API across all applications).

[0272] The central intelligence layer can include a number of machine-learned models. For example a respective machine-learned model (e.g., a model) can be provided for each application and managed by the central intelligence layer. In other implementations, two or more applications can share a single machine-learned model. For example, in some implementations, the central intelligence layer can provide a single model (e.g., a single model) for all of the applications. In some implementations, the central intelligence layer is included within or otherwise implemented by an operating system of the computing system 1000.

[0273] The central intelligence layer can communicate with a central device data layer. The central device data layer can be a centralized repository of data for the computing system 1000. The central device data layer may communicate with a number of other components of the computing device, such as, for example, one or more sensors, a context manager, a device state component, and/or additional components. In some implementations, the central device data layer can communicate with each device component using an API (e.g., a private API).

[0274] Figure 19B depicts a block diagram of an example computing system 1150 that performs document-to-dialogue audio conversion according to example embodiments of the present disclosure. In particular, the example computing system 1150 can include one or more computing devices 1152 that can be utilized to obtain, and/or generate, one or more datasets that can be processed by a sensor processing system 1160 and/or an output determination system 1180 to feedback to a user that can provide information on features in the one or more obtained datasets. The one or more datasets can include image data, text data, audio data, multimodal data, latent encoding data, etc. The one or more datasets may be obtained via one or more sensors associated with the one or more computing devices 1152 (e.g., one or more sensors in the computing device 1152). Additionally and/or alternatively, the one or more datasets can be stored data and/or retrieved data (e.g., data retrieved from a web resource). For example, images, text, and/or other content items may be interacted with by a user. The interacted with content items can then be utilized to generate one or more determinations.

[0275] The one or more computing devices 1152 can obtain, and/or generate, one or more datasets based on image capture, sensor tracking, data storage retrieval, content download (e.g., downloading an image or other content item via the internet from a web resource), and/or via one or more other techniques. The one or more datasets can be processed with a sensor processing system 1160. The sensor processing system 1160 may perform one or more processing techniques using one or more machine-learned models, one or more search engines, and/or one or more other processing techniques. The one or more processing techniques can be performed in any combination and/or individually. The one or more processing techniques can be performed in series and/or in parallel. In particular, the one or more datasets can be processed with a context determination block 1162, which may determine a context associated with one or more content items. The context determination block 1162 may identify and/or process metadata, user profile data (e.g., preferences, user search history, user browsing history, user purchase history, and/or user input data), previous interaction data, global trend data, location data, time data, and/or other data to determine a particular context associated with the user. The context can be associated with an event, a determined trend, a particular action, a particular type of data, a particular environment, and/or another context associated with the user and/or the retrieved or obtained data.

[0276] The sensor processing system 1160 may include an image preprocessing block 1164. The image preprocessing block 1164 may be utilized to adjust one or more values of an obtained and/or received image to prepare the image to be processed by one or more machine-learned models and/or one or more search engines 1174. The image preprocessing block 1164 may resize the image, adjust saturation values, adjust resolution, strip and/or add metadata, and/or perform one or more other operations.

[0277] In some implementations, the sensor processing system 1160 can include one or more machine-learned models, which may include a detection model 1166, a segmentation model 1168, a classification model 1170, an embedding model 1172, and/or one or more other machine-learned models. For example, the sensor processing system 1160 may include one or more detection models 66 that can be utilized to detect particular features in the processed dataset. In particular, one or more images can be processed with the one or more detection models 66 to generate one or more bounding boxes associated with detected features in the one or more images.

[0278] Additionally and/or alternatively, one or more segmentation models 1168 can be utilized to segment one or more portions of the dataset from the one or more datasets. For example, the one or more segmentation models 1168 may utilize one or more segmentation masks (e.g., one or more segmentation masks manually generated and/or generated based on the one or more bounding boxes) to segment a portion of an image, a portion of an audio file, and/or a portion of text. The segmentation may include isolating one or more detected objects and/or removing one or more detected objects from an image.

[0279] The one or more classification models 1170 can be utilized to process image data, text data, audio data, latent encoding data, multimodal data, and/or other data to generate one or more classifications. The one or more classification models 1170 can include one or more image classification models, one or more object classification models, one or more text classification models, one or more audio classification models, and/or one or more other classification models. The one or more classification models 1170 can process data to determine one or more classifications.

[0280] In some implementations, data may be processed with one or more embedding models 1172 to generate one or more embeddings. For example, one or more images can be processed with the one or more embedding models 1172 to generate one or more image embeddings in an embedding space. The one or more image embeddings may be associated with one or more image features of the one or more images. In some implementations, the one or more embedding models 1172 may be configured to process multimodal data to generate multimodal embeddings. The one or more embeddings can be utilized for classification, search, and/or learning embedding space distributions.

[0281] The sensor processing system 1160 may include one or more search engines 1174 that can be utilized to perform one or more searches. The one or more search engines 1174 may crawl one or more databases (e.g., one or more local databases, one or more global databases, one or more private databases, one or more public databases, one or more specialized databases, and/or one or more general databases) to determine one or more search results. The one or more search engines 1174 may perform feature matching, text based search, embedding based search (e.g., k-nearest neighbor search), metadata based search, multimodal search, web resource search, image search, text search, and/or application search. [0282] Additionally and/or alternatively, the sensor processing system 1160 may include one or more multimodal processing blocks 1176, which can be utilized to aid in the processing of multimodal data. The one or more multimodal processing blocks 1176 may include generating a multimodal query and/or a multimodal embedding to be processed by one or more machine-learned models and/or one or more search engines 1174.

[0283] The output(s) of the sensor processing system 1160 can then be processed with an output determination system 1180 to determine one or more outputs to provide to a user. The output determination system 1180 may include heuristic based determinations, machine- learned model based determinations, user selection based determinations, and/or context based determinations.

[0284] The output determination system 1180 may determine how and/or where to provide the one or more search results in a search results interface 1182. Additionally and/or alternatively, the output determination system 1180 may determine how and/or where to provide the one or more machine-learned model outputs in a machine-learned model output interface 1184. In some implementations, the one or more search results and/or the one or more machine-learned model outputs may be provided for display via one or more user interface elements. The one or more user interface elements may be overlayed over displayed data. For example, one or more detection indicators may be overlaid over detected objects in a viewfinder. The one or more user interface elements may be selectable to perform one or more additional searches and/or one or more additional machine-learned model processes. In some implementations, the user interface elements may be provided as specialized user interface elements for specific applications and/or may be provided uniformly across different applications. The one or more user interface elements can include pop-up displays, interface overlays, interface tiles and/or chips, carousel interfaces, audio feedback, animations, interactive widgets, and/or other user interface elements.

[0285] Additionally and/or alternatively, data associated with the output(s) of the sensor processing system 1160 may be utilized to generate and/or provide an augmented- reality experience and/or a virtual -reality experience 1186. For example, the one or more obtained datasets may be processed to generate one or more augmented-reality rendering assets and/or one or more virtual-reality rendering assets, which can then be utilized to provide an augmented-reality experience and/or a virtual-reality experience 1186 to a user. The augmented-reality experience may render information associated with an environment into the respective environment. Alternatively and/or additionally, objects related to the processed dataset(s) may be rendered into the user environment and/or a virtual environment. Rendering dataset generation may include training one or more neural radiance field models to learn a three-dimensional representation for one or more objects.

[0286] In some implementations, one or more action prompts 1188 may be determined based on the output(s) of the sensor processing system 1160. For example, a search prompt, a purchase prompt, a generate prompt, a reservation prompt, a call prompt, a redirect prompt, and/or one or more other prompts may be determined to be associated with the output(s) of the sensor processing system 1160. The one or more action prompts 1188 may then be provided to the user via one or more selectable user interface elements. In response to a selection of the one or more selectable user interface elements, a respective action of the respective action prompt may be performed (e.g., a search may be performed, a purchase application programming interface may be utilized, and/or another application may be opened).

[0287] In some implementations, the one or more datasets and/or the output(s) of the sensor processing system 1160 may be processed with one or more generative models 1190 to generate a model -generated content item that can then be provided to a user. The generation may be prompted based on a user selection and/or may be automatically performed (e.g., automatically performed based on one or more conditions, which may be associated with a threshold amount of search results not being identified).

[0288] The one or more generative models 1190 can include language models (e.g., large language models and/or vision language models), image generation models (e.g., text- to-image generation models and/or image augmentation models), audio generation models, video generation models, graph generation models, and/or other data generation models (e.g., other content generation models). The one or more generative models 1190 can include one or more transformer models, one or more convolutional neural networks, one or more recurrent neural networks, one or more feedforward neural networks, one or more generative adversarial networks, one or more self-attention models, one or more embedding models, one or more encoders, one or more decoders, and/or one or more other models. In some implementations, the one or more generative models 1190 can include one or more autoregressive models (e.g., a machine-learned model trained to generate predictive values based on previous behavior data) and/or one or more diffusion models (e.g., a machine- learned model trained to generate predicted data based on generating and processing distribution data associated with the input data).

[0289] The one or more generative models 1190 can be trained to process input data and generate model-generated content items, which may include a plurality of predicted words, pixels, signals, and/or other data. The model -generated content items may include novel content items that are not the same as any pre-existing work. The one or more generative models 1190 can leverage learned representations, sequences, and/or probability distributions to generate the content items, which may include phrases, storylines, settings, objects, characters, beats, lyrics, and/or other aspects that are not included in pre-existing content items.

[0290] The one or more generative models 1190 may include a vision language model.

[0291] The vision language model can be trained, tuned, and/or configured to process image data and/or text data to generate a natural language output. The vision language model may leverage a pre-trained large language model (e.g., a large autoregressive language model) with one or more encoders (e.g., one or more image encoders and/or one or more text encoders) to provide detailed natural language outputs that emulate natural language composed by a human.

[0292] The vision language model may be utilized for zero-shot image classification, few shot image classification, image captioning, multimodal query distillation, multimodal question and answering, and/or may be tuned and/or trained for a plurality of different tasks. The vision language model can perform visual question answering, image caption generation, feature detection (e.g., content monitoring (e.g., for inappropriate content)), object detection, scene recognition, and/or other tasks.

[0293] The vision language model may leverage a pre-trained language model that may then be tuned for multimodality. Training and/or tuning of the vision language model can include image-text matching, masked-language modeling, multimodal fusing with cross attention, contrastive learning, prefix language model training, and/or other training techniques. For example, the vision language model may be trained to process an image to generate predicted text that is similar to ground truth text data (e.g., a ground truth caption for the image). In some implementations, the vision language model may be trained to replace masked tokens of a natural language template with textual tokens descriptive of features depicted in an input image. Alternatively and/or additionally, the training, tuning, and/or model inference may include multi-layer concatenation of visual and textual embedding features. In some implementations, the vision language model may be trained and/or tuned via jointly learning image embedding and text embedding generation, which may include training and/or tuning a system to map embeddings to a joint feature embedding space that maps text features and image features into a shared embedding space. The joint training may include image-text pair parallel embedding and/or may include triplet training. In some implementations, the images may be utilized and/or processed as prefixes to the language model.

[0294] The output determination system 1180 may process the one or more datasets and/or the output(s) of the sensor processing system 1160 with a data augmentation block 1192 to generate augmented data. For example, one or more images can be processed with the data augmentation block 1192 to generate one or more augmented images. The data augmentation can include data correction, data cropping, the removal of one or more features, the addition of one or more features, a resolution adjustment, a lighting adjustment, a saturation adjustment, and/or other augmentation.

[0295] In some implementations, the one or more datasets and/or the output(s) of the sensor processing system 1160 may be stored based on a data storage block 1194 determination.

[0296] The output(s) of the output determination system 1180 can then be provided to a user via one or more output components of the user computing device 1152. For example, one or more user interface elements associated with the one or more outputs can be provided for display via a visual display of the user computing device 1152.

[0297] The processes may be performed iteratively and/or continuously. One or more user inputs to the provided user interface elements may condition and/or affect successive processing loops.

[0298] The technology discussed herein makes reference to servers, databases, software applications, and other computer-based systems, as well as actions taken and information sent to and from such systems. The inherent flexibility of computer-based systems allows for a great variety of possible configurations, combinations, and divisions of tasks and functionality between and among components. For instance, processes discussed herein can be implemented using a single device or component or multiple devices or components working in combination. Databases and applications can be implemented on a single system or distributed across multiple systems. Distributed components can operate sequentially or in parallel.

[0299] While the present subject matter has been described in detail with respect to various specific example embodiments thereof, each example is provided by way of explanation, not limitation of the disclosure. Those skilled in the art, upon attaining an understanding of the foregoing, can readily produce alterations to, variations of, and equivalents to such embodiments. Accordingly, the subject disclosure does not preclude inclusion of such modifications, variations or additions to the present subject matter as would be readily apparent to one of ordinary skill in the art. For instance, features illustrated or described as part of one embodiment can be used with another embodiment to yield a still further embodiment. Thus, it is intended that the present disclosure cover such alterations, variations, and equivalents.

[0300] Aspects of the disclosure have been described in terms of illustrative embodiments thereof. Any and all features in the following claims can be combined or rearranged in any way possible, including combinations of claims not explicitly enumerated in combination together, as the example claim dependencies listed herein should not be read as limiting the scope of possible combinations of features disclosed herein. Accordingly, the scope of the present disclosure is by way of example rather than by way of limitation, and the subject disclosure does not preclude inclusion of such modifications, variations or additions to the present subject matter as would be readily apparent to one of ordinary skill in the art. Moreover, terms are described herein using lists of example elements joined by conjunctions such as “and,” “or,” “but,” etc. It should be understood that such conjunctions are provided for explanatory purposes only. Clauses and other sequences of items joined by a particular conjunction such as “or,” for example, can refer to “and/or,” “at least one of’, “any combination of’ example elements listed therein, etc. Terms such as “based on” should be understood as “based at least in part on.”

[0301] The term “can” should be understood as referring to a possibility of a feature in various implementations and not as prescribing an ability that is necessarily present in every implementation. For example, the phrase “X can perform Y” should be understood as indicating that, in various implementations, X has the potential to be configured to perform Y, and not as indicating that in every instance X must always be able to perform Y. It should be understood that, in various implementations, X might be unable to perform Y and remain within the scope of the present disclosure.

[0302] The term “may” should be understood as referring to a possibility of a feature in various implementations and not as prescribing an ability that is necessarily present in every implementation. For example, the phrase “X may perform Y” should be understood as indicating that, in various implementations, X has the potential to be configured to perform Y, and not as indicating that in every instance X must always be able to perform Y. It should be understood that, in various implementations, X might be unable to perform Y and remain within the scope of the present disclosure.

Claims

WHAT IS CLAIMED IS:

1. A computing system for document-to-dialogue audio conversion, the system comprising: one or more processors; and one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations, the operations comprising: obtaining a document; processing the document with a generative language model to generate a model-generated script, wherein the model -generated script comprises a plurality of interleaved dialogue turns between two or more speakers; processing the document and the model -generated script with the generative language model to generate a revised script; and processing the revised script with a text-to-speech model to generate audio data, wherein the audio data is descriptive of the revised script being output in audio form by a first speaker voice and a second speaker voice; and providing the audio data for playback.

2. The system of claim 1, wherein the operations further comprise: evaluating, with a machine-learned evaluation model, the model-generated script to generate an evaluation model output; and wherein the revised script is generated based in part on the evaluation model output.

3. The system of claim 1, wherein the revised script comprises a first set of dialogue associated with a first speaker and a second set of dialogue associated with a second speaker.

4. The system of claim 3, wherein the operations further comprise: processing the revised script and the audio data to determine the second speaker voice speaks a dialogue line associated with the first speaker; in response to determining the second speaker voice speaks the dialogue line associated with the first speaker, generating an additional text-to-speech prompt; and processing the revised script and the additional text-to-speech prompt with the text-to- speech model to generate second audio data.

5. The system of claim 3, wherein the first set of dialogue is associated with a host role, and wherein the second set of dialogue is associated with a technical expert role.

6. The system of claim 3, wherein the first set of dialogue comprises questions, and wherein the second set of dialogue comprises responses to the questions, wherein the responses are generated based on content of the research paper.

7. The system of claim 1, wherein processing the document with the generative language model to generate the model-generated script comprises: generating semantic understanding data based on content and structure of the academic paper; generating a first set of questions; generating a first set of responses responsive to the first set of questions based on the semantic understanding data; and interweaving the first set of questions and the first set of responses to generate a chain of dialogue.

8. The system of claim 7, wherein the first set of questions is generated based on determining the academic paper is associated with a particular information type based on the semantic understanding data.

9. The system of claim 1, wherein the revised script comprises a second plurality of interleaved dialogue turns between the two or more speakers.

10. The system of claim 1, wherein the revised script comprises a second modelgenerated script that differs from the model -generated script based on the generative language model being conditioned on the evaluation model output.

11. A computer-implemented method for document-to-dialogue audio conversion, the method comprising: obtaining, by a computing system comprising one or more processors, a document, wherein the document comprises an academic paper; processing, by the computing system, the document with a generative language model to generate a model -generated script, wherein the model -generated script comprises a plurality of interleaved dialogue turns between two or more speakers; evaluating, by the computing system and with a machine-learned evaluation model, the model -generated script to generate an evaluation model output; processing, by the computing system, the document, the model -generated script, and the evaluation model output with the generative language model to generate a revised script; and processing, by the computing system, the revised script with a text-to-speech model to generate audio data, wherein the audio data is descriptive of the revised script being output in audio form by a first speaker voice and a second speaker voice.

12. The method of claim 11, further comprising: determining, by the computing system, the academic paper is older than a threshold age; in response to determining the academic paper is older than the threshold age, obtaining an age-disclaimer prompt; and processing, by the computing system, the age-disclaimer prompt, the revised script, the model -generated evaluation output, and the document with the generative language model to generate an augmented script.

13. The method of claim 12, wherein the augmented script comprises an augmented version of the revised script, wherein the revised script is augmented to include one or more disclaimers on a year of publication of the academic paper.

14. The method of claim 11, wherein evaluating, with the machine-learned evaluation model, the model-generated script to generate the evaluation model output comprises: processing the model -generated script and the document with the machine-learned evaluation model to determine the model -generated script comprises one or more factual inaccuracies; and generating, with the machine-learned evaluation model, the evaluation model output that comprises one or more facts from the document.

15. The method of claim 11, wherein evaluating, with the machine-learned evaluation model, the model-generated script to generate the evaluation model output comprises: processing the model -generated script and the document with the machine-learned evaluation model to determine the model -generated script comprises a first person perspective when referring to the document; and generating, with the machine-learned evaluation model, the evaluation model output that comprises instructions to include dialogue from a perspective of a reader.

16. The method of claim 11, wherein the generative language model comprises an autoregressive language model, and wherein the machine-learned evaluation model comprises a natural language processing model that comprises a plurality of classifiers and a prompt generation model.

17. One or more non-transitory computer-readable media that collectively store instructions that, when executed by one or more computing devices, cause the one or more computing devices to perform operations, the operations comprising: obtaining a document, wherein the document comprises an academic paper; processing the document with a generative language model to generate a modelgenerated script, wherein the model-generated script comprises a plurality of interleaved dialogue turns between two or more speakers; processing the model -generated script with a machine-learned evaluation model to generate an evaluation model output; generating a revision prompt based on the evaluation model output; processing the document, the revision prompt, and the evaluation model output with the generative language model to generate a revised script; and processing the revised script with a text-to-speech model to generate audio data, wherein the audio data is descriptive of the revised script being output in audio form by a first speaker voice and a second speaker voice.

18. The one or more non-transitory computer-readable media of claim 17, wherein the text-to-speech model comprises a text-to- speech autoencoder.

19. The one or more non-transitory computer-readable media of claim 17, wherein the text-to-speech model comprises a transformer model.

20. The one or more non-transitory computer-readable media of claim 17, wherein the operations further comprise: obtaining a second document, wherein the second document comprises a second academic paper; concatenating the document and the second document to generate a combined document; and processing the combined document with the generative language model to generate a multi-document script.