JP7689541B2 - Information processing method, model training method, device, equipment, medium, and program product - Google Patents

Description

The present disclosure relates to the field of computer technology, in particular to the field of artificial intelligence and speech technology, and specifically to information processing methods, model training methods, devices, equipment, media, and program products.

With the development of natural language processing technology, machine learning models can be used in the field of smart dialogue, and the dialogue model can respond based on the sentences entered by the user, achieving the effect of dialogue with the user.

Currently, the dialogue model has low dialogue accuracy and poor dialogue quality.

The present disclosure provides an information processing method, a model training method, an apparatus, a device, a medium, and a program product.

According to one aspect of the present disclosure, there is provided an information processing method, the method comprising:
obtaining an initial dialogue;
inputting the initial dialogue sentence into a trained dialogue model to obtain a target response sentence;
The dialogue model is a model obtained by training based on a corrected response sample sentence, a second candidate response sample sentence, and a recall response sample sentence, and an initial dialogue sample sentence is input into an initial dialogue model to obtain a plurality of candidate response sample sentences, the second candidate response sample sentence being any one of the plurality of candidate response sample sentences, the corrected response sample sentence being a sentence obtained by correcting a first response sample sentence among the candidate response sample sentences, and the recall response sample sentence being a sample sentence other than the initial dialogue sample sentence and the plurality of candidate response sample sentences among the training sample sentences.

According to another aspect of the present disclosure, there is provided a model training method, the method comprising:
obtaining an initial dialogue sample sentence;
inputting the initial dialogue sample sentence into an initial dialogue model to obtain a plurality of candidate reply sample sentences;
modifying a first candidate reply sample sentence from the plurality of candidate reply sample sentences to obtain a modified reply sample sentence;
training the initial dialogue model based on the revised response sample sentence, a second candidate response sample sentence of the plurality of candidate response sample sentences, and a recall response sample sentence to obtain a dialogue model;
The recall response sample sentences are sample sentences other than the initial dialogue sample sentences and the plurality of candidate response sample sentences, among the training sample sentences.

According to another aspect of the present disclosure, there is provided an information processing device, the device comprising:
an acquisition module for acquiring an initial dialogue;
an input module for inputting the initial dialogue sentence into a trained dialogue model to obtain a target response sentence;
The dialogue model is a model obtained by training based on a corrected response sample sentence, a second candidate response sample sentence, and a recall response sample sentence, and an initial dialogue sample sentence is input into an initial dialogue model to obtain a plurality of candidate response sample sentences, the second candidate response sample sentence being any one of the plurality of candidate response sample sentences, the corrected response sample sentence being a sentence obtained by correcting a first response sample sentence among the candidate response sample sentences, and the recall response sample sentence being a sample sentence other than the initial dialogue sample sentence and the plurality of candidate response sample sentences among the training sample sentences.

According to another aspect of the present disclosure, there is provided a model training apparatus, the apparatus comprising:
a sentence acquisition module for acquiring an initial dialogue sample sentence;
a sentence input module for inputting the initial dialogue sample sentence into an initial dialogue model to obtain a plurality of candidate reply sample sentences;
a correction module that corrects a first candidate reply sample sentence from the plurality of candidate reply sample sentences to obtain a corrected reply sample sentence;
a training module for training the initial dialogue model based on the revised response sample sentence, a second candidate response sample sentence among the plurality of candidate response sample sentences, and a recall response sample sentence to obtain a dialogue model;
The recall response sample sentences are sample sentences other than the initial dialogue sample sentences and the plurality of candidate response sample sentences, among the training sample sentences.

According to another aspect of the present disclosure, there is provided an electronic device, the electronic device comprising:
At least one processor;
a memory communicatively coupled to the at least one processor;
The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor such that the at least one processor performs the above-described methods.

According to another aspect of the present disclosure, a non-transitory computer-readable storage medium having computer instructions stored thereon causes a computer to perform the method described above.

According to another aspect of the present disclosure, there is provided a computer program that, when executed by a processor, implements the steps of the above method.

In some embodiments of the present disclosure, a dialogue model is obtained by training based on the corrected response sample sentence, the second candidate response sample sentence, and the recall response sample sentence, and an initial dialogue sample sentence is input into the initial dialogue model to obtain a plurality of candidate response sample sentences, the second candidate response sample sentence being one of the plurality of candidate response sample sentences, the corrected response sample sentence being a high quality dialogue sentence obtained by correcting the first response sample sentence among the candidate response sample sentences, and the recall response sample sentence being a sample sentence other than the initial dialogue sample sentence and the plurality of candidate response sample sentences among the training sample sentences, and a dialogue model with high dialogue accuracy is obtained by continuing to train the initial dialogue model on the corrected response sample sentence, the second candidate response sample sentence, and the recall response sample sentence, and the initial dialogue sentence is input into the dialogue model to obtain a target response sentence with high dialogue quality.

It should be understood that the contents described in this section are not intended to identify key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will be readily understood from the following description.

The drawings are used for a better understanding of the present technical solution, and are not intended to limit the present disclosure.
1 is a schematic flowchart of an information processing method provided by the first embodiment of the present disclosure; 1 is a schematic flowchart of a model training method provided by Example 2 of the present disclosure; 11 is a flowchart of an information processing method provided by Example 3 of the present disclosure. FIG. 1 is a schematic configuration diagram of an information processing device provided by an exemplary embodiment of the present disclosure. FIG. 1 is a schematic configuration diagram of a model training apparatus provided by an exemplary embodiment of the present disclosure. FIG. 1 is a schematic block diagram of an exemplary electronic device for implementing embodiments of the present disclosure.

Below, exemplary embodiments of the present disclosure are described in conjunction with the drawings, and various details of the embodiments of the present disclosure are included therein for ease of understanding, and should be considered as merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for clarity and conciseness, the following description omits descriptions of well-known functions and structures.

In addition, in the technical solution disclosed herein, the collection, storage, use, processing, transmission, provision, disclosure and other processing of relevant user personal information shall all comply with the provisions of relevant laws and regulations and shall not violate public order and morals.

Artificial intelligence is a discipline that studies how computers can simulate certain human thought processes and intelligent behaviors (learning, reasoning, thinking, planning, etc.), and includes both hardware and software level technologies. AI hardware technology generally includes sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, and other technologies. AI software technology mainly includes computer vision technology, voice recognition technology, natural language processing technology, and several other aspects such as machine learning/deep learning, big data processing technology, and knowledge graph technology.

With the development of natural language processing technology, machine learning models can be used in the field of smart dialogue, and the dialogue model can respond based on the sentences entered by the user, achieving the effect of dialogue with the user.

In the field of dialogue systems, large-scale dialogue models trained on social media review data are emerging one after another. However, due to a discrepancy between social media review scenes and actual human dialogue scenes, the model's generation ability is poor.

A generative dialogue model generates multiple candidate replies during inference and then evaluates and sorts the replies using a generation score. However, sorting methods based on the generation score cannot effectively bring high-quality replies to the front of the queue.

Currently, the dialogue model has low dialogue accuracy and poor dialogue quality.

In response to the above technical problems, in some embodiments of the present disclosure, a dialogue model is obtained by training based on the corrected response sample sentence, the second candidate response sample sentence, and the recall response sample sentence, and an initial dialogue sample sentence is input into the initial dialogue model to obtain a plurality of candidate response sample sentences, the second candidate response sample sentence being one of the plurality of candidate response sample sentences, the corrected response sample sentence being a high quality dialogue sentence obtained by correcting the first response sample sentence among the candidate response sample sentences, and the recall response sample sentence being a sample sentence other than the initial dialogue sample sentence and the plurality of candidate response sample sentences among the training sample sentences, and a dialogue model with high dialogue accuracy is obtained by continuing to train the initial dialogue model on the corrected response sample sentence, the second candidate response sample sentence, and the recall response sample sentence, and the initial dialogue sentence is input into the dialogue model to obtain a target response sentence with high dialogue quality.

The technical solutions provided by each embodiment of the present disclosure are described in detail below with reference to the drawings.

Figure 1 is a schematic flowchart of an information processing method provided by the first embodiment of the present disclosure. As shown in Figure 1, the method includes the following steps S101 to S102.

S101: An initial dialogue is obtained.
S102, input the initial dialogue sentence into a trained dialogue model to obtain a target response sentence.
The dialogue model is a model obtained by training based on the corrected response sample sentence, the second candidate response sample sentence and the recall response sample sentence, and the initial dialogue sample sentence is input into the initial dialogue model to obtain a plurality of candidate response sample sentences, the second candidate response sample sentence being any one of the plurality of candidate response sample sentences, the corrected response sample sentence being a sentence obtained by correcting a first reply sample sentence among the candidate response sample sentences, and the recall response sample sentence being any other sample sentence among the training sample sentences excluding the initial dialogue sample sentence and the plurality of candidate response sample sentences.

In this embodiment, the entity that executes the above method may be a server or a terminal device.

When the execution entity of the above method is a server, the implementation form of the server is not limited. For example, the server may be a server device such as a general-purpose server, a cloud server, a cloud host, or a virtual center. The configuration of the server mainly includes a processor, a hard disk, a memory, a system bus, etc., and a type of general-purpose computer architecture.

When the execution entity of the above method is a terminal device, the implementation form of the terminal device is not limited. The terminal device includes, but is not limited to, a personal computer, a tablet computer, a smartphone, or a smart wearable device.

In this embodiment, training is performed based on the corrected response sample sentence, the second candidate response sample sentence, and the recall response sample sentence to obtain a dialogue model, and the initial dialogue sample sentence is input into the initial dialogue model to obtain multiple candidate response sample sentences, the second candidate response sample sentence is one of the multiple candidate response sample sentences, the corrected response sample sentence is a high-quality dialogue obtained by correcting the first response sample sentence among the candidate response sample sentences, and the recall response sample sentence is a sample sentence other than the initial dialogue sample sentence and the multiple candidate response sample sentences among the training sample sentences, and by continuing to train the initial dialogue model for the corrected response sample sentence, the second candidate response sample sentence, and the recall response sample sentence, a dialogue model with high dialogue accuracy is obtained, an initial dialogue sentence is obtained, and the initial dialogue sentence is input into the dialogue model to obtain a target response sentence with high dialogue quality.

Below, we will explain the technical proposals disclosed in this disclosure according to application scenarios.

Application scenario 1: The smartphone responds to an initial dialogue sentence "How is the weather today?" input by the user via voice, the smartphone uploads the initial dialogue sentence to the server, the server inputs the initial dialogue sentence into a trained dialogue model to obtain the target response sentence "It's sunny today", the server transmits the target response sentence down to the smartphone, and the smartphone plays the target response sentence "It's sunny today" via voice.

Application scenario 2: The smartphone responds to an initial dialogue sentence "How is the weather today?" input by the user through voice, and inputs the initial dialogue sentence into the locally integrated dialogue model to obtain a target response sentence "It's sunny today", and the smartphone plays the target response sentence "It's sunny today" through voice.

Before using a dialogue model, we need to train an initial dialogue model to obtain a dialogue model. The process of training a dialogue model is described below.

Figure 2 is a schematic flowchart of a model training method provided by the second embodiment of the present disclosure. As shown in Figure 2, the method includes the following steps S201 to S204.

S201: Obtain initial dialogue sample sentences.

S202: Input the initial dialogue sample sentence into the initial dialogue model to obtain multiple candidate response sample sentences.

S203: Modify a first candidate response sample sentence from among the plurality of candidate response sample sentences to obtain a modified response sample sentence.

S204, training an initial dialogue model based on the revised response sample sentence, a second candidate response sample sentence among the plurality of candidate response sample sentences, and the recall response sample sentence to obtain a dialogue model.
The recall response sample sentences are the training sample sentences other than the initial dialogue sample sentences and the plurality of candidate response sample sentences.

The training device for training the above dialogue model may be any type of computing device, and the embodiments of the present disclosure are not limited thereto.

Note that the initial dialogue model may be a trained model, and the accuracy of the initial dialogue model may be low, resulting in poor dialogue quality when using the initial dialogue model.

Obtain an initial dialogue sample sentence, input the initial dialogue sample sentence into an initial dialogue model to obtain a revised response sample sentence; modify a first candidate response sample sentence from the plurality of candidate response sample sentences to obtain a revised response sample sentence; randomly select a second candidate response sample sentence from the plurality of candidate response sample sentences; and select a recall response sample sentence from the other sample sentences in the training sample sentences excluding the initial dialogue sample sentence and the plurality of candidate response sample sentences. The revised response sample sentence, the second candidate response sample sentence, and the recall response sample sentence constitute one training dataset. The above steps are repeated to obtain a training dataset for model training.

In order to increase the coverage of the dataset, the initial dialogue sample sentences will be taken from datasets in as many different fields as possible, such as news, social media, literature, and live-action dialogue.

In the above embodiment, a first candidate response sample sentence among a plurality of candidate response sample sentences is modified to obtain a modified response sample sentence. For example, an operation such as copying, correcting, or creating is performed on the first candidate response sample sentence to obtain a modified response sample sentence.

For example, in response to an operation of inputting an initial dialogue sample sentence in the labeling interface, the initial dialogue sample sentence "It rains every day and makes me feel bad" is obtained, and the initial dialogue sample sentence is input into the initial dialogue model to obtain a number of candidate response sample sentences "Music and chocolate go well with rainy days", "Rainy days are perfect for sleeping", "I feel bad too because no one will join me", "Rainy days are nice", "Me too! I don't like rainy days", "Yes, it's a problem that I can't go out", and "Yes, I don't like rainy days either".

A first candidate response sample sentence "Music and chocolate go well with rainy days" from among the multiple candidate response sample sentences is modified to obtain a modified response sample sentence "I think music and chocolate go well with rainy days", a second candidate response sample sentence "Rainy days are perfect for sleeping" is randomly selected from the multiple candidate response sample sentences, and a recall response sample sentence "It's sunny today" is selected from the other sample sentences excluding the initial dialogue sample sentence and the multiple candidate response sample sentences from the training sample sentences. The modified response sample sentence "Music and chocolate go well with rainy days", the second candidate response sample sentence "Rainy days are perfect for sleeping", and the recall response sample sentence "It's sunny today" constitute one training dataset.

In the above embodiment, the initial dialogue model is trained based on the corrected response sample sentence, the second candidate response sample sentence and the recall response sample sentence among the multiple candidate response sample sentences, to obtain a dialogue model. In one possible embodiment, the corrected response sample sentence, the second candidate response sample sentence and the recall response sample sentence are input into an initial sentence generation model of the initial dialogue model to obtain the actual response sentence, the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence and the probability of the recall response sample sentence, and the initial sentence generation model and the initial sentence determination model of the initial dialogue model are jointly trained based on the actual response sentence, the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence and the probability of the recall response sample sentence to obtain a dialogue model.

In one embodiment, an initial sentence generation model and an initial sentence determination model of an initial dialogue model are jointly trained based on the actual response sentence, the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence, and the probability of the recall response sample sentence to obtain a dialogue model. A loss function is determined based on the actual response sentence and the corrected response sample sentence, and based on the loss function, the probability of the corrected response sample sentence is greater than the probability of the second candidate response sample sentence, the probability of the corrected response sample sentence is greater than the probability of the recall response sample sentence, and the probability of the second candidate response sample sentence is greater than the probability of the recall response sample sentence, as training targets. The initial sentence generation model and the initial sentence determination model are jointly trained to obtain a dialogue model.

In conjunction with the above description of each embodiment, FIG. 3 is a flowchart of an information processing method provided by the third embodiment of the present disclosure. As shown in FIG. 3, the method includes the following steps S301 to S304.

S301: The terminal device responds to a voice input operation and obtains an initial dialogue.

S302: The terminal device transmits the initial dialogue to the server.

S303: The server receives the initial dialogue, inputs the initial dialogue into the dialogue model to obtain a target response sentence, and transmits the target response sentence down to the terminal device.

S304: The terminal device receives the target response sentence and plays the target response sentence aloud.

In this embodiment, the implementation form of the server is not limited. For example, the server may be a server device such as a general-purpose server, a cloud server, a cloud host, or a virtual center. The configuration of the server mainly includes a processor, a hard disk, a memory, a system bus, etc., and a type of general-purpose computer architecture.

In this embodiment, the implementation form of the terminal device is not limited. The terminal device may include, but is not limited to, a personal computer, a tablet computer, a smartphone, or a smart wearable device.

The implementation form of each step in this embodiment can be referred to the explanation of the above embodiments, and the explanation will be omitted in this embodiment, while at the same time, this embodiment can obtain the beneficial effects of the parts corresponding to each of the above embodiments.

FIG. 4 is a schematic diagram of an information processing device 40 provided by an exemplary embodiment of the present disclosure. The information processing device 40 includes an acquisition module 41 and an input module 42.

The acquisition module 41 acquires the initial dialogue.

The input module 42 inputs the initial dialogue sentence into the trained dialogue model to obtain the target response sentence.
The dialogue model is a model obtained by training based on the corrected response sample sentence, the second candidate response sample sentence and the recall response sample sentence, and the initial dialogue sample sentence is input into the initial dialogue model to obtain a plurality of candidate response sample sentences, the second candidate response sample sentence being any one of the plurality of candidate response sample sentences, the corrected response sample sentence being a sentence obtained by correcting a first reply sample sentence among the candidate response sample sentences, and the recall response sample sentence being any other sample sentence among the training sample sentences excluding the initial dialogue sample sentence and the plurality of candidate response sample sentences.

Optionally, the input module 42 inputs the initial dialogue sentence into the trained dialogue model to obtain the target response sentence,
Within the dialogue model, inputting the initial dialogue sentence into a sentence generation model of the dialogue model to obtain a plurality of candidate response sentences and a probability of each candidate response sentence;
A plurality of candidate response sentences and the probability of each candidate response sentence are input to a sentence determination model of the dialogue model to obtain a target response sentence.

Optionally, the input module 42 inputs a plurality of candidate response sentences and the probability of each candidate response sentence into a sentence determination model of the dialogue model to obtain a target response sentence,
A plurality of candidate reply sentences and the probability of each candidate reply sentence are input to a sentence determination model, and a target reply sentence with the highest probability is selected from the plurality of candidate reply sentences.

5 is a schematic diagram of a model training device 50 provided by an exemplary embodiment of the present disclosure. The model training device 50 includes a sentence acquisition module 51, a sentence input module 52, a correction module 53, and a training module 54.
The sentence acquisition module 51 acquires an initial dialogue sample sentence,
The sentence input module 52 inputs the initial dialogue sample sentence into the initial dialogue model to obtain a plurality of candidate reply sample sentences;
The correction module 53 corrects a first candidate reply sample sentence of the plurality of candidate reply sample sentences to obtain a corrected reply sample sentence;
a training module 54 training an initial dialogue model based on the revised response sample sentence, a second candidate response sample sentence among the plurality of candidate response sample sentences, and the recall response sample sentence to obtain a dialogue model;
The recall response sample sentences are sample sentences other than the initial dialogue sample sentences and the plurality of candidate response sample sentences among the training sample sentences.

Optionally, the training module 54 trains an initial dialogue model based on the revised response sample sentence, the second candidate response sample sentence of the plurality of candidate response sample sentences, and the recall response sample sentence to obtain a dialogue model,
inputting the corrected response sample sentence, the second candidate response sample sentence and the recall response sample sentence into an initial sentence generation model of the initial dialogue model to obtain the actual response sentence, the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence and the probability of the recall response sample sentence;
An initial sentence generation model and an initial sentence determination model of the initial dialogue model are jointly trained based on the actual response sentence, the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence, and the probability of the recall response sample sentence to obtain a dialogue model.

Optionally, the training module 54 jointly trains the initial sentence generation model and the initial sentence determination model of the initial dialogue model according to the actual response sentence, the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence, and the probability of the recall response sample sentence to obtain a dialogue model:
determining a loss function based on the actual response sentence and the sample corrected response sentence;
Based on the loss function, the probability of the corrected response sample sentence is greater than the probability of the second candidate response sample sentence, the probability of the corrected response sample sentence is greater than the probability of the recall response sample sentence, and the probability of the second candidate response sample sentence is greater than the probability of the recall response sample sentence, and the initial sentence generation model and the initial sentence determination model are jointly trained to obtain a dialogue model.

The specific method for executing the operations of each module of the device in the above embodiment has already been described in detail in the embodiment relating to the method, and will not be described in detail here.

According to an embodiment of the present disclosure, the present disclosure further provides an electronic device and a readable storage medium.
According to an embodiment of the present disclosure, the present disclosure further provides a computer program, which, when executed by a processor, realizes the information processing method or the model training method provided by the present disclosure.

FIG. 6 is a schematic block diagram of an exemplary electronic device 600 for implementing embodiments of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, mobile phones, smartphones, wearable devices, and other similar computing devices. The components, their connections and relationships, and their functions shown herein are merely examples and are not intended to limit the description herein and/or the implementation of the present disclosure as sought.

As shown in FIG. 6, the electronic device 600 includes a computing unit 601 that can perform various appropriate operations and processes according to a computer program stored in a read-only memory (ROM) 602 or loaded from a storage unit 608 into a random access memory (RAM) 603. The RAM 603 may also store various programs and data required for the operation of the electronic device 600. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input/output (I/O) interface 605 is also connected to the bus 604.

The components of the electronic device 600 are connected to an I/O interface 605, which includes an input unit 606 such as a keyboard, a mouse, etc., an output unit 607 such as various types of displays, speakers, etc., a storage unit 608 such as a magnetic disk, an optical disk, etc., and a communication unit 609 such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 enables the electronic device 600 to exchange information/data with other devices via a computer network such as the Internet and/or various telecommunication networks.

The computing unit 601 may be various general-purpose and/or dedicated processing components having processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphic processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units that execute machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 executes each of the methods and processes described above, such as the information processing method and the model training method. For example, in some embodiments, the information processing method and the model training method can be realized as a computer software program tangibly included in a machine-readable medium such as the storage unit 608. In some embodiments, some or all of the computer program can be loaded and/or installed in the electronic device 600 via the ROM 602 and/or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the information processing method and the model training method described above may be performed. Alternatively, in other embodiments, the computation unit 601 may be configured in any other suitable manner (e.g., via firmware) to perform the information processing method and the model training method.

Various embodiments of the systems and techniques described herein above may be implemented in digital electronic circuitry systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs that may be executed and/or interpreted by a programmable system that includes at least one programmable processor, which may be an application specific or general purpose programmable processor, and that may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

Program codes for carrying out the methods of the present disclosure can be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, a special purpose computer, or other programmable data processing apparatus such that, when executed by the processor or controller, the functions/operations specified in the flowcharts and/or block diagrams are performed. The program codes may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine, partially on a remote machine, or entirely on a remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain or store a program for use by or in combination with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the above content. More specific examples of machine-readable storage media include one or more line-based electrical connections, portable computer disks, hard disks, random access memories (RAMs), read-only memories (ROMs), erasable programmable read-only memories (EPROMs or flash memories), optical fibers, portable compact disk read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the above content.

To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user, and a keyboard and pointing device (e.g., a mouse or trackball) by which the user can provide input to the computer. Other types of devices can also provide interaction with a user, for example, the feedback provided to the user can be any form of sensing feedback (e.g., vision feedback, auditory feedback, or haptic feedback) and can receive input from the user in any form (including acoustic, speech, or haptic input).

The systems and techniques described herein may be implemented in a computing system that includes a back-end component (e.g., a data server), or a computing system that includes a middleware component (e.g., an application server), or a computing system that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with embodiments of the systems and techniques described herein), or any combination of such back-end, middleware, and front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), the Internet, and a blockchain network.

The computer system may include a client and a server. The client and server are generally remote from each other and typically interact with each other via a communication network. The relationship between the client and the server is generated by computer programs running on corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also called a cloud computing server or cloud host, which is a host product in a cloud computing service system and solves the deficiencies of difficult management and weak business scalability that exist in traditional physical hosts and VPS services (abbreviated as "Virtual Private Server", or "VPS"). The server may be a server of a distributed system or a server incorporating blockchain.

It should be understood that steps can be rearranged, added, or removed using the various types of flows shown above. For example, the steps described in this disclosure may be performed in parallel, sequentially, or in a different order, but are not limited herein as long as the technical solutions disclosed in this disclosure can achieve the desired results.

The above specific embodiments do not limit the scope of protection of the present disclosure. It should be understood that those skilled in the art may make various modifications, combinations, subcombinations, and substitutions according to design requirements and other factors. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present disclosure should be included within the scope of protection of the present disclosure.

Claims (13)

An information processing method executed by an information processing device,
obtaining an initial dialogue;
inputting the initial dialogue sentence into a trained dialogue model to obtain a target response sentence;
The dialogue model is a model obtained by training based on a corrected response sample sentence, a second candidate response sample sentence, and a recall response sample sentence, and an initial dialogue sample sentence is input to an initial dialogue model to obtain a plurality of candidate response sample sentences, the second candidate response sample sentence being any one of the plurality of candidate response sample sentences, the corrected response sample sentence being a sentence obtained by correcting a first reply sample sentence among the candidate response sample sentences, and the recall response sample sentence being a sample sentence other than the initial dialogue sample sentence and the plurality of candidate response sample sentences among the training sample sentences, and the corrected response sample sentence, the second candidate response sample sentence, and the recall response sample sentence are inputting sample sentences into an initial sentence generation model of the initial dialogue model to obtain probabilities of corrected response sample sentences, probabilities of second candidate response sample sentences and probabilities of recall response sample sentences; generating actual response sentences according to the probabilities of the corrected response sample sentences, the probabilities of the second candidate response sample sentences and the probabilities of the recall response sample sentences generated by the initial sentence generation model through an initial sentence determination model; and acquiring the dialogue model by jointly training the initial sentence generation model and the initial sentence determination model of the initial dialogue model according to the probabilities of the actual response sentences, the probabilities of the corrected response sample sentences, the probabilities of the second candidate response sample sentences and the probabilities of the recall response sample sentences;
23. An information processing method comprising: inputting the initial dialogue sentence into a trained dialogue model to obtain a target response sentence,
within the dialogue model, inputting the initial dialogue sentence into a sentence generation model of the dialogue model to obtain a plurality of candidate response sentences and a probability of each of the candidate response sentences;
inputting the plurality of candidate response sentences and a probability of each of the candidate response sentences into a sentence determination model of the dialogue model to obtain a target response sentence;
2. The information processing method according to claim 1, inputting the plurality of candidate response sentences and the probabilities of each of the candidate response sentences into a sentence determination model of the dialogue model to obtain a target response sentence,
inputting the plurality of candidate response sentences and a probability of each of the candidate response sentences into the sentence determination model, and selecting a most probable target response sentence from the plurality of candidate response sentences.
3. The information processing method according to claim 2. A model training method executed by a model training device, comprising:
obtaining an initial dialogue sample sentence;
inputting the initial dialogue sample sentence into an initial dialogue model to obtain a plurality of candidate reply sample sentences;
modifying a first candidate reply sample sentence from the plurality of candidate reply sample sentences to obtain a modified reply sample sentence;
training the initial dialogue model based on the revised response sample sentence, a second candidate response sample sentence of the plurality of candidate response sample sentences, and a recall response sample sentence to obtain a dialogue model;
the recall response sample sentence is a sample sentence other than the initial dialogue sample sentence and the plurality of candidate response sample sentences among the training sample sentences,
training the initial dialogue model based on the revised response sample sentence, a second candidate response sample sentence of the plurality of candidate response sample sentences, and a recall response sample sentence to obtain a dialogue model,
inputting the corrected response sample sentence, the second candidate response sample sentence and the recall response sample sentence into an initial sentence generation model of the initial dialogue model to obtain the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence and the probability of the recall response sample sentence; and then generating an actual response sentence according to the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence and the probability of the recall response sample sentence generated by the initial sentence generation model through an initial sentence determination model;
and co- training the initial sentence generation model and the initial sentence determination model of the initial dialogue model based on the probabilities of the actual response sentences, the corrected response sample sentences, the second candidate response sample sentences, and the recall response sample sentences to obtain the dialogue model.
A model training method comprising: a step of jointly training the initial sentence generation model and the initial sentence determination model of the initial dialogue model based on the actual response sentence, the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence, and the probability of the recall response sample sentence to obtain the dialogue model,
determining a loss function based on the actual response sentence and the corrected response sample sentence;
and co-training the initial sentence generation model and the initial sentence determination model to obtain the dialogue model, based on the loss function, with training targets that the probability of the corrected response sample sentence is greater than the probability of the second candidate response sample sentence, the probability of the corrected response sample sentence is greater than the probability of the recall response sample sentence, and the probability of the second candidate response sample sentence is greater than the probability of the recall response sample sentence.
5. The method of claim 4, wherein the model training method is An information processing device,
an acquisition module for acquiring an initial dialogue;
an input module for inputting the initial dialogue sentence into a trained dialogue model to obtain a target response sentence;
The dialogue model is a model obtained by training based on a corrected response sample sentence, a second candidate response sample sentence, and a recall response sample sentence, and an initial dialogue sample sentence is input to an initial dialogue model to obtain a plurality of candidate response sample sentences, the second candidate response sample sentence being any one of the plurality of candidate response sample sentences, the corrected response sample sentence being a sentence obtained by correcting a first reply sample sentence among the candidate response sample sentences, and the recall response sample sentence being a sample sentence other than the initial dialogue sample sentence and the plurality of candidate response sample sentences among the training sample sentences, and the corrected response sample sentence, the second candidate response sample sentence, and the recall response sample sentence are inputting sample sentences into an initial sentence generation model of the initial dialogue model to obtain probabilities of corrected response sample sentences, probabilities of second candidate response sample sentences and probabilities of recall response sample sentences; generating actual response sentences according to the probabilities of the corrected response sample sentences, the probabilities of the second candidate response sample sentences and the probabilities of the recall response sample sentences generated by the initial sentence generation model through an initial sentence determination model; and acquiring the dialogue model by jointly training the initial sentence generation model and the initial sentence determination model of the initial dialogue model according to the probabilities of the actual response sentences, the probabilities of the corrected response sample sentences, the probabilities of the second candidate response sample sentences and the probabilities of the recall response sample sentences;
23. An information processing apparatus comprising: When the input module inputs the initial dialogue sentence into a trained dialogue model to obtain a target response sentence,
Within the dialogue model, inputting the initial dialogue sentence into a sentence generation model of the dialogue model to obtain a plurality of candidate response sentences and a probability of each of the candidate response sentences;
inputting the plurality of candidate response sentences and the probability of each of the candidate response sentences into a sentence determination model of the dialogue model to obtain a target response sentence;
7. The information processing apparatus according to claim 6, the input module inputs the plurality of candidate response sentences and the probabilities of each of the candidate response sentences into a sentence determination model of the dialogue model to obtain a target response sentence,
inputting the plurality of candidate response sentences and a probability of each of the candidate response sentences into the sentence determination model, and selecting a target response sentence with the highest probability from the plurality of candidate response sentences;
8. The information processing apparatus according to claim 7, A model training device,
a sentence acquisition module for acquiring an initial dialogue sample sentence;
a sentence input module for inputting the initial dialogue sample sentence into an initial dialogue model to obtain a plurality of candidate reply sample sentences;
a correction module that corrects a first candidate reply sample sentence from the plurality of candidate reply sample sentences to obtain a corrected reply sample sentence;
a training module for training the initial dialogue model based on the revised response sample sentence, a second candidate response sample sentence among the plurality of candidate response sample sentences, and a recall response sample sentence to obtain a dialogue model;
the recall response sample sentence is a sample sentence other than the initial dialogue sample sentence and the plurality of candidate response sample sentences among the training sample sentences,
when the training module trains the initial dialogue model based on the revised response sample sentence, a second candidate response sample sentence among the plurality of candidate response sample sentences, and a recall response sample sentence to obtain a dialogue model,
inputting the corrected response sample sentence, the second candidate response sample sentence and the recall response sample sentence into an initial sentence generation model of the initial dialogue model to obtain the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence and the probability of the recall response sample sentence; and then generating an actual response sentence according to the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence and the probability of the recall response sample sentence generated by the initial sentence generation model using an initial sentence determination model ;
jointly training the initial sentence generation model and the initial sentence determination model of the initial dialogue model based on the actual response sentence, the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence, and the probability of the recall response sample sentence to obtain the dialogue model;
A model training device characterized by: the training module jointly trains the initial sentence generation model and the initial sentence determination model of the initial dialogue model according to the actual response sentence, the probability of the corrected response sample sentence, the probability of the second candidate response sample sentence, and the probability of the recall response sample sentence to obtain the dialogue model;
determining a loss function based on the actual response sentence and the sample corrected response sentence;
Based on the loss function, a probability of the corrected response sample sentence is greater than a probability of the second candidate response sample sentence, a probability of the corrected response sample sentence is greater than a probability of the recall response sample sentence, and a probability of the second candidate response sample sentence is greater than a probability of the recall response sample sentence, using these as training targets, to jointly train the initial sentence generation model and the initial sentence determination model to obtain the dialogue model.
10. The model training device according to claim 9. An electronic device,
At least one processor;
a memory communicatively coupled to the at least one processor;
The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor such that the at least one processor performs a method according to any one of claims 1 to 3 or 4 or 5.
1. An electronic device comprising: A non-transitory computer-readable storage medium having computer instructions stored thereon, comprising:
The computer instructions cause a computer to carry out a method according to any one of claims 1 to 3 or 4 or 5.
A non-transitory computer-readable storage medium comprising: A computer program comprising:
The computer program, when executed by a processor, implements the steps of the method according to any one of claims 1 to 3, 4 or 5.
A computer program comprising:

Images