JP7797264B2 - Information processing device, information processing method, and information processing program - Google Patents

Description

This application relates to an information processing device, an information processing method, and an information processing program.

Advertising distribution via the Internet has traditionally been popular. For example, in traditional advertising distribution, advertising content such as still images, moving images, and text promoting companies, products, etc. are displayed in advertising spaces set up in designated locations on web pages. In traditional advertising distribution, when such advertising content is clicked, the user is redirected to a landing page, such as the advertiser's web page or content introducing the advertising target.

In addition, with conventional advertising delivery, targeted delivery may be used to selectively deliver advertisements that correspond to user information such as user preferences, gender, age, address, and occupation, in order to increase advertising effectiveness.

JP 2010-136332 A

However, traditional ad delivery aims to achieve specific results through advertising, and new ad delivery methods are required that are not bound by existing ad delivery methods.

This application has been made in light of the above, and aims to provide an information processing device, information processing method, and information processing program that can realize new forms of ad delivery without being bound by existing methods of ad delivery.

The information processing device according to the present application has a selection unit and a distribution unit. The selection unit selects multiple advertisements from advertisements submitted by advertisers other than the requesting advertiser as materials for generating a composite advertisement for conveying information about the requesting advertiser requesting the generation of the composite advertisement. The distribution unit distributes the composite advertisement generated by combining the advertisements selected by the selection unit.

According to one aspect of the embodiment, new ad delivery can be realized without being bound by existing ad delivery methods.

FIG. 1 is a diagram showing an overview of a method for distributing a composite advertisement according to an embodiment. FIG. 2 is a diagram showing an outline of an advertisement selection method (part 1) according to an embodiment. FIG. 3 is a diagram showing an outline of an advertisement selection method (part 2) according to the embodiment. FIG. 4 is a diagram showing an outline of an advertisement selection method (part 3) according to the embodiment. FIG. 5 is a diagram showing an overview of an advanced version of the method for delivering a composite advertisement according to the embodiment. FIG. 6 is a diagram illustrating an example of the configuration of an information processing apparatus according to the embodiment. FIG. 7 is a diagram illustrating an example of advertisement information according to the embodiment. FIG. 8 is a diagram illustrating an example of distribution result information according to the embodiment. FIG. 9 is a diagram illustrating an example of business type correspondence information according to the embodiment. FIG. 10 is a flowchart illustrating an example of a procedure (part 1) for distributing a composite advertisement according to the embodiment. FIG. 11 is a flowchart illustrating an example of a procedure (part 2) for delivering a composite advertisement according to the embodiment. FIG. 12 is a flowchart illustrating an example of a procedure (part 3) for delivering a composite advertisement according to the embodiment. FIG. 13 is a hardware configuration diagram illustrating an example of a computer that realizes the functions of the information processing apparatus according to the embodiment.

Below, detailed explanations are given of modes for implementing the information processing device, information processing method, and information processing program (hereinafter referred to as "embodiments") according to the present application, with reference to the drawings. Note that the information processing device, information processing method, and information processing program according to the present application are not limited to the multiple embodiments described below. Furthermore, the multiple embodiments described below can be combined as appropriate to the extent that the processing content is not contradictory. Furthermore, the same components in the multiple embodiments described below will be assigned the same reference numerals, and redundant explanations will be omitted.

[1. Overview of information processing]
(1-1. Synthetic Ad Delivery Method)
An example of information processing according to an embodiment will be described with reference to FIG. 1. FIG. 1 is a diagram illustrating an overview of a method for distributing a composite advertisement according to an embodiment. Note that FIG. 1 describes an example of information processing in the case where, when a web page related to a predetermined service is provided in response to a request from a user (e.g., user U1) who is a service user, advertising content is displayed in an advertising space provided at a predetermined position on the web page. Note that the information processing according to the embodiment is not limited to the case where the advertising medium is a web page, and can be similarly applied to the case where the advertising medium is email.

In the example shown in FIG. 1, the information processing system SYS has a user terminal 10 and an information processing device 100. The user terminal 10 and the information processing device 100 are connected to a network N (see FIG. 6, for example) via a wired or wireless connection. The user terminal 10 and the information processing device 100 can communicate with each other via the network N. Note that the example shown in FIG. 1 is not particularly limited, and the information processing system SYS may include more user terminals 10 and information processing devices 100 than those shown in FIG. 1.

The user terminal 10 shown in FIG. 1 is an information processing terminal used by user U1, who is a user of various services provided by the information processing device 100. A typical example of the user terminal 10 is a smartphone, but it may also be implemented as a tablet terminal, a notebook PC (Personal Computer), a desktop PC, a mobile phone, a PDA (Personal Digital Assistant), etc. In the example of FIG. 1, a smartphone is shown as the user terminal 10. Note that in the following description, the user terminal 10 may be referred to as user U1. In other words, user U1 can be read as user terminal 10. Note that in the following description, a user of various services provided by the information processing device 100 may be referred to simply as a "user."

The user terminal 10 also has a communication unit for connecting to the network N via a wireless communication network such as LTE (Long Term Evolution), 4G (4th Generation), or 5G (5th Generation: fifth generation mobile communication system), or via short-range wireless communication such as Bluetooth (registered trademark) or wireless LAN (Local Area Network). A user U operates the user terminal 10 having the communication unit to access the information processing device 100 and view information on various services.

The user terminal 10 can also display information for using various services provided by the information processing device 100 using a web browser or application. When the user terminal 10 receives control information from the information processing device 100 to realize information display processing using a web browser or application, the user terminal 10 realizes the display processing in accordance with the received control information.

The information processing device 100 shown in FIG. 1 is an information processing device that provides various services to a user U. The information processing device 100 is typically a server device, but may also be realized by a mainframe, workstation, or the like. When the information processing device 100 is realized by a server device, it may be realized by a single server, or may be realized by a cloud system in which multiple server devices and multiple storage devices operate in cooperation with each other.

The information processing device 100 may also function as a distribution device that distributes control information to user terminals 10 used by users of various services. Here, the control information is written in, for example, a script language such as JavaScript (registered trademark) or a style sheet language such as CSS (Cascading Style Sheets). The application itself distributed from the information processing device 100 may also be considered to be control information.

The various services provided by the information processing device 100 may also include API (Application Programming Interface) services corresponding to various applications, as well as various online services. The various online services may include Q&A services, Internet connections, search services, SNS (Social Networking Services), e-commerce services, electronic payment services, online games, online banking services, online trading services, hotel reservation services, ticket reservation services, video distribution services, music distribution services, news distribution services, map information services, route search services, route guidance services, line information services, operation information services, and weather information services.

The information processing device 100 also functions as a device for delivering advertising content. The information processing device 100 manages advertising content submitted by advertisers and manages information related to the results of advertisement delivery (delivery history). For example, the information processing device 100 displays advertising content submitted by advertisers in advertising spaces provided on web pages of various services provided in response to user requests.

Advertisement bidding for ad spaces provided on web pages is conducted using a predetermined bidding method, such as RTB (Real-Time Bidding). The advertisement of the advertiser who wins the bid for the ad space is then displayed in the ad space. As a result, advertisements from advertisers who lack financial resources and are not price competitive have fewer opportunities for users to encounter the advertisement. In light of this situation, the information processing device 100 according to the embodiment displays a composite advertisement that combines multiple advertisements from different advertisers in the ad space, with the aim of improving ad delivery from the advertiser's perspective. Below, the information processing according to the embodiment will be described in detail, following the flow that is expected when this information processing is performed.

For example, the user terminal 10 sends an access request to a specified service to the information processing device 100 in accordance with the operation of the user U (step S11).

When the information processing device 100 receives a request to access a predetermined service from the user terminal 10, it selects a first advertisement (hereinafter referred to as the "primary advertisement") that is the primary advertisement and a second advertisement (hereinafter referred to as the "secondary advertisement") that is the secondary advertisement (step S12). The primary advertisement is a part of the advertisements that make up the composite advertisement, and is the advertisement that accepts the secondary advertisement as part of the advertisement. The secondary advertisement is a part of the advertisements that make up the composite advertisement, and is the advertisement that is inserted into part of the primary advertisement.

For example, the information processing device 100 selects a primary advertisement and a secondary advertisement from among multiple advertisements to be combined, based on advertisement type information preset for each advertisement. The advertisement type is set by the advertiser when submitting the advertisement. As shown in FIG. 1, advertisement types include three types: "normal advertisement," "parasitic-accepting advertisement," and "parasitic-wanted advertisement." "Normal advertisement" is a type selected by an advertiser when the advertiser wishes to distribute the advertisement as a conventional advertisement rather than as a combined advertisement. "Parasitic-accepting advertisement" is a type selected by an advertiser when the advertiser wishes to allow the advertisement of another advertiser to be combined as a combined advertisement. An advertisement set as "parasitic-accepting advertisement" is treated as a primary advertisement. "Parasitic-wanted advertisement" is a type selected by an advertiser when the advertiser wishes to combine the advertisement of another advertiser as a combined advertisement. An advertisement set as "parasitic-wanted advertisement" is treated as a secondary advertisement.

The information processing device 100 selects one primary advertisement from among the advertisements whose advertisement type is "parasitic-accepting advertisement." The information processing device 100 also selects one secondary advertisement from among the advertisements whose advertisement type is "parasitic-wanting advertisement." Figure 1 shows an example in which "Advertisement A" is selected as the primary advertisement and "Advertisement B" is selected as the secondary advertisement. Note that the information victory device 100 may select multiple secondary advertisements when generating a composite advertisement, meaning that at least one secondary advertisement is selected to be combined with the primary advertisement to generate a composite advertisement.

After selecting the advertisements, the information processing device 100 generates a composite advertisement by combining a portion of the primary advertisement with the secondary advertisement (step S13). In the composite advertisement generated by the information processing device 100, the primary advertisement is displayed larger than the second advertisement, which is the secondary advertisement. Note that the layout and display size of the secondary advertisement relative to the primary advertisement can be determined arbitrarily, provided that at least the advertising target of the primary advertisement is recognizable to the user. Figure 1 shows an example of a composite advertisement AD_a in which a portion of advertisement A selected as the primary advertisement is combined with advertisement B selected as the secondary advertisement.

After generating the composite advertisement, the information processing device 100 distributes the generated composite advertisement to the user terminal 10 (step S14). The user terminal 10 displays the composite advertisement distributed from the information processing device 100 in an advertisement space provided on a web page corresponding to a specified service.

As described above, the information processing device 100 selects a primary advertisement and a secondary advertisement, generates a composite advertisement by combining a portion of the primary advertisement with the secondary advertisement, and distributes the generated composite advertisement. In this way, the information processing device 100 can improve the way advertisements are distributed from the advertiser's perspective. In other words, by distributing an advertisement from an advertiser with weak price competitiveness as a secondary advertisement by combining it with a portion of a price-competitive primary advertisement, the information processing device 100 can increase the likelihood of achieving a predetermined result for the secondary advertisement.

The information processing device 100 may also analyze the delivery results of the composite advertisement and select the primary advertisement and the secondary advertisement while avoiding combinations of advertisements for which the index value indicating the advertising effectiveness of the composite advertisement is below a predetermined standard. The information processing device 100 may also analyze the delivery results of the composite advertisement and select the primary advertisement and the secondary advertisement while prioritizing combinations of advertisements for which the index value indicating the advertising effectiveness of the composite advertisement is equal to or greater than a predetermined standard. This allows the information processing device 100 to further increase the likelihood of achieving predetermined results for the secondary advertisement. When selecting advertisements, the information processing device 100 may select advertisements as an optimization problem, in addition to using neural networks or vector search. For example, the information processing device 100 may perform optimization by combining constraints such as the budget of the primary advertisement and the secondary advertisement, or the click-through rate (CTR) performance.

In addition, the information processing device 100 may select the primary and secondary advertisements by taking into account information regarding combinations of advertisements that should be avoided. For example, an example of a combination of advertisements that should be avoided is a combination of an automobile advertisement and an alcoholic beverage advertisement. This allows the information processing device 100 to avoid inappropriate advertising language when delivering a composite advertisement.

In addition, the information processing device 100 may select the primary and secondary advertisements by taking into account information indicating the competitive relationships between advertisers. This allows the information processing device 100 to take care, when delivering a composite advertisement, to ensure that the market share of the transaction subject, such as a product or service offered by the advertiser of the primary advertisement, is not eroded.

Furthermore, when a composite advertisement is selected (clicked) at the distribution destination, the information processing device 100 may determine the amount to be charged to each advertiser associated with the composite advertisement. For example, the information processing device 100 may charge only the advertiser of the primary advertisement. In this case, the information processing device 100 may charge the amount to be charged to the advertiser of the primary advertisement lower than the cost per click corresponding to a "normal" advertisement. Furthermore, the information processing device 100 may charge the advertiser corresponding to the clicked advertisement of the primary advertisement and the secondary advertisements that make up the composite advertisement. In this case, the information processing device 100 may prorate the total amount of the cost per click of the primary advertisement and the cost per click of the secondary advertisement according to the display size of the composite advertisement, and determine the prorated amount as the amount to be charged. This allows the information processing device 100 to implement charging processing tailored to the composite advertisement.

(1-2. Ad selection method (part 1))
Furthermore, the information processing device 100 is not particularly limited to selecting advertisements based on advertisement types preset by advertisers. For example, the information processing device 100 may input information about advertisements that are candidates for the primary advertisement and information about advertisements that are candidates for the secondary advertisement, and perform advertisement selection based on the output of a trained model that outputs an estimated value of advertising effectiveness obtained from a combined advertisement that combines the information about the primary advertisement and the secondary advertisement. An outline of an advertisement selection method (part 1) according to an embodiment will be described below with reference to FIG. 2. FIG. 2 is a diagram illustrating an outline of the advertisement selection method (part 1) according to an embodiment.

As shown in FIG. 2, the information processing device 100 selects a primary advertisement and a secondary advertisement based on the output of the advertising effectiveness estimation model M-1, which is a trained model (step S21). The information processing device 100 trains the model to learn the compatibility characteristics when combining primary advertisements and secondary advertisements based on the delivery results of the combined advertisement. As a result, the information processing device 100 inputs information about the primary advertisement and secondary advertisements to be combined, and generates the advertising effectiveness estimation model M-1, which is a trained model that outputs an estimated value of the advertising effectiveness obtained by the combined advertisement obtained by combining them. The learning process for generating the advertising effectiveness estimation model M-1 is described below.

For example, the information processing device 100 trains a model to learn the correspondence between information about the primary advertisement and the secondary advertisement, the approval result for the acceptance of the secondary advertisement by the advertiser of the primary advertisement, and a score based on the advertising effectiveness of the combined advertisement.

The information about the primary and secondary advertisements may be any information related to the advertisement. For example, the information about the advertisement may be a title displayed as text related to the advertisement, a description summarizing the content of the advertising page (web page) on which the advertisement is displayed, or information about the content of the advertisement, such as images or videos used as advertising materials. Furthermore, if the advertiser is a company, the information about the advertisement may be information such as capital and number of employees, or if the advertiser is a listed company, it may be stock price information.

The approval result for the primary advertiser's acceptance of the secondary advertisement is a history of information indicating whether the primary advertiser approved or denied the acceptance of the secondary advertisement as part of the primary advertisement. The advertising effectiveness of the composite advertisement is determined based on a predetermined indicator of advertising effectiveness. The predetermined indicator of advertising effectiveness can be any indicator, such as cost per acquisition (CPA), cost per click (CPC), click-through rate (CTR), number of conversions, conversion rate (CVR), or page view (PV).

For example, the administrator of the information processing device 100 pre-sets a score that reflects the result of approval of the acceptance of the secondary advertisement by the advertiser of the primary advertisement and an index value that indicates the advertising effectiveness of the combined advertisement. Specifically, the administrator sets the score so that the score increases as the acceptance of the secondary advertisement is approved and the index value increases, and decreases when the acceptance of the secondary advertisement is rejected. The administrator then associates the set score with a combination of information about the primary advertisement and the secondary advertisement, and creates a learning data sample DT.

The information processing device 100 uses the created training data samples to train a model to understand the correspondence between information about the primary and secondary advertisements, the approval result for the acceptance of the secondary advertisement by the primary advertisement's advertiser, and a score based on the advertising effectiveness of the composite advertisement. The information processing device 100 can use any network, such as a neural network, as a training model to train the model to understand the correspondence. Furthermore, when using a neural network as a training model, the information processing device 100 performs a training process to adjust the values of the weights (connection coefficients) that are taken into account when scores (values) are propagated between nodes in each layer that make up the neural network. For example, the information processing device 100 uses any method, such as backpropagation, to perform a process to optimize (correct) the parameters (connection coefficients) so as to reduce the error between the output from the training model and the correct answer (correct answer data) corresponding to the input. When the information processing device 100 uses the backpropagation method, it executes a learning process to minimize a predetermined loss function, thereby correcting the parameters (connection coefficients) to reduce the error between the output from the learning model and the correct answer (correct data) corresponding to the input.

Through the above-described learning process, the information processing device 100 can generate an advertising effectiveness estimation model M-1, which is a trained model that inputs information about the primary and secondary advertisements to be combined and outputs an estimated value of the advertising effectiveness obtained by combining them into a combined advertisement.

The information processing device 100 inputs all combinations of primary and secondary advertisements to be combined into the advertising effectiveness estimation model M-1, and determines the primary and secondary advertisements to be selected as the combination targets based on the estimated advertising effectiveness output for each combination from the advertising effectiveness estimation model M-1. For example, the information processing device 100 determines the primary and secondary advertisements that constitute the combination with the largest estimated advertising effectiveness as the combination targets. Figure 2 shows an example in which "Advertisement A" is selected as the primary advertisement and "Advertisement B" is selected as the secondary advertisement.

Then, based on the output from the advertising effectiveness estimation model M-1, the information processing device 100 generates a composite advertisement using the primary advertisement and the secondary advertisement selected as the subjects of combination (step S22). For example, the information processing device 100 generates a composite advertisement by combining a portion of the primary advertisement with the secondary advertisement using the primary advertisement and the secondary advertisement selected as the subjects of combination. Figure 2 shows an example of a composite advertisement AD_a in which a portion of advertisement A selected as the primary advertisement is combined with advertisement B selected as the secondary advertisement.

As described above, the information processing device 100 can estimate the advertising effectiveness obtained from a combined advertisement that combines a primary advertisement and a secondary advertisement selected as the targets of combination, thereby enabling the delivery of a highly effective combined advertisement.

(1-3. Ad selection method (part 2))
Furthermore, the information processing device 100 may input information about the advertisers of the primary advertisement and the secondary advertisements to be combined, and perform advertisement selection based on the output of a trained model that vectorizes the primary advertisement and the secondary advertisements, respectively. An outline of the advertisement selection method (part 2) according to the embodiment will be described below with reference to FIG. 3. FIG. 3 is a diagram illustrating an outline of the advertisement selection method (part 2) according to the embodiment.

As shown in FIG. 3, the information processing device 100 selects a primary advertisement and a secondary advertisement based on the output of the advertising vector output model M-2, which is a trained model (step S31). The information processing device 100 trains a model that vectorizes each advertisement based on information about the advertisers of each advertisement that constitutes the composite advertisement and the delivery results of the composite advertisement. As a result, the information processing device 100 inputs information about the primary advertisement and secondary advertisement to be combined, and generates the advertising vector output model M-2, which is a trained model that outputs a feature vector obtained by vectorizing the features of the information about the primary advertisement and a feature vector obtained by vectorizing the features of the information about the secondary advertisement. The learning process for generating the advertising vector output model M-2 is described below.

For example, the administrator of the information processing device 100 sets a score based on information about the advertisers of the primary and secondary advertisements, the approval result for the primary advertisement's acceptance of the secondary advertisement by the advertiser, and the advertising effectiveness of the combined advertisement.

The information about the advertisers of the primary and secondary advertisements may be any information related to the advertisement. For example, if the advertiser is a company, it may be information such as capital and number of employees, and if it is a listed company, it may be information about the stock price. Note that information about the advertisers of the primary and secondary advertisements may also be used when setting the score. In this case, the information about the advertisement may be information about the content of the advertisement, such as the title displayed as text related to the advertisement, a description summarizing the content of the advertising page (web page) on which the advertisement is displayed, or images and videos used as advertising materials.

The approval result for the primary advertiser's acceptance of the secondary advertisement is a history of information indicating whether the primary advertiser approved or denied the acceptance of the secondary advertisement as part of the primary advertisement. The advertising effectiveness of the composite advertisement is determined based on a predetermined indicator of advertising effectiveness. The predetermined indicator of advertising effectiveness can be any indicator, such as cost per acquisition (CPA), cost per click (CPC), click-through rate (CTR), number of conversions, conversion rate (CVR), or page view (PV).

For example, the administrator of the information processing device 100 pre-sets a score that reflects the result of approval of the acceptance of the secondary advertisement by the advertiser of the primary advertisement and an index value that indicates the advertising effectiveness of the combined advertisement. Specifically, the administrator sets the score so that the score increases as the acceptance of the secondary advertisement is approved and the index value increases, and decreases when the acceptance of the secondary advertisement is rejected. The administrator then associates the set score with a combination of information about the advertisers of the primary advertisement and the secondary advertisement, and creates a learning data sample DT.

The information processing device 100 uses the created training data sample DT to train a model so that the larger the score set for a combination of information about the advertisers of the primary and secondary advertisements, the smaller the similarity between the feature vector corresponding to the advertiser of the primary advertisement and the feature vector corresponding to the advertiser of the secondary advertisement is output as a vector. Specifically, the information processing device 100 trains a model that outputs each feature vector so that the larger the set score in the training data sample DT, the smaller the cosine similarity between the feature vector corresponding to the primary advertisement and the feature vector corresponding to the secondary advertisement.

For example, when the information processing device 100 uses an arbitrary network such as a neural network as a training model, it performs a training process to adjust the values of weights (connection coefficients) that are taken into account when scores (values) are propagated between nodes in each layer that make up the neural network. For example, the information processing device 100 uses an arbitrary method such as backpropagation to optimize (correct) the parameters (connection coefficients) so as to reduce the error between the output from the training model and the correct answer (correct data) corresponding to the input. In other words, the information processing device 100 optimizes the parameters (connection coefficients) so that the cosine similarity between the feature vectors corresponding to the primary advertisement and the feature vectors corresponding to the secondary advertisement decreases as the set score in the training data sample DT increases. Note that when using the backpropagation method, the information processing device 100 can correct the parameters (connection coefficients) so as to reduce the error between the output from the training model and the correct answer (correct data) corresponding to the input by performing a training process to minimize a predetermined loss function.

Through the above-described learning process, the information processing device 100 can generate an advertising vector output model M-2, which is a trained model that vectorizes each advertisement based on information about the advertisers of each advertisement that constitutes the composite advertisement and the delivery results of the composite advertisement.

For each combination of primary and secondary advertisements to be combined, the information processing device 100 inputs information about the advertiser into the advertising vector output model M-2, and determines the primary and secondary advertisements to be selected as combination targets based on the similarity between the feature vectors output from the advertising vector output model M-2. For example, the information processing device 100 determines the primary and secondary advertisements with the smallest similarity between their feature vectors as combination targets. Figure 3 shows an example in which "Advertisement A" is selected as the primary advertisement and "Advertisement B" is selected as the secondary advertisement.

Then, the information processing device 100 generates a composite advertisement using the primary advertisement and the secondary advertisement selected as the subjects of compositing (step S32). For example, the information processing device 100 generates a composite advertisement by using the primary advertisement and the secondary advertisement selected as the subjects of compositing to combine a portion of the primary advertisement with the secondary advertisement. Figure 3 shows an example of a composite advertisement AD_a in which advertisement B selected as the secondary advertisement is combined with a portion of advertisement A selected as the primary advertisement.

In addition, the information processing device 100 may train the model so that the closer the scale of the advertisers, the smaller the similarity between the vectors corresponding to each advertisement that makes up the composite advertisement. For example, when creating the training data sample DT, the administrator of the information processing device 100 reflects the scale of the advertiser companies (capital, number of employees, stock price, etc.) in the set score. Specifically, the administrator adjusts the set score so that the closer the scale of the advertiser companies, the smaller the set score. Then, the information processing device 100 performs the training process using the training data sample DT in which the scale of the advertisers is reflected.

In addition, the information processing device 100 may train the model so that the more the composite advertisement is selected (clicked), the smaller the similarity between the vectors corresponding to each advertisement that makes up the composite advertisement. For example, when creating the training data sample DT, the administrator of the information processing device 100 reflects an index value indicating the advertising effectiveness (such as click rate or conversion rate) in the set score. Specifically, the administrator adjusts the set score so that the larger the index value of the advertising effectiveness, the smaller the set score. Then, the information processing device 100 performs the training process using the training data sample DT that reflects the size of the advertiser.

As described above, the information processing device 100 can verify the similarity between the advertiser of the primary advertisement selected as the combination target and the advertiser of the secondary advertisement, thereby realizing the delivery of a highly effective combined advertisement. In other words, by selecting advertisements from advertisers of different scales as the combination target, the information processing device 100 can increase the appeal of the combined advertisement to users, and high advertising effectiveness can be expected. Furthermore, the information processing device 100 can expect high advertising effectiveness by selecting advertisements as the combination target that match a combination of advertisers with high advertising effectiveness.

(1-4. Ad Selection Method (Part 3))
Furthermore, the information processing device 100 may automatically select a primary advertisement and a secondary advertisement from among a plurality of advertisements to be combined in accordance with a predetermined selection rule, regardless of the advertisement type described in the advertisement selection method (part 1). An outline of the advertisement selection method (part 3) according to the embodiment will be described below with reference to FIG. 4. FIG. 4 is a diagram illustrating an outline of the advertisement selection method (part 3) according to the embodiment.

As shown in FIG. 4, the information processing device 100 automatically selects a primary advertisement and a secondary advertisement from among multiple advertisements to be combined, based on the delivery results of the combined advertisement generated by combining a portion of the primary advertisement with a secondary advertisement (step S41).

Figure 4 illustrates an example of a composite advertisement delivery result in which the click-through rate, which indicates the advertising effectiveness of a composite advertisement consisting of primary advertisement A and secondary advertisement B, is greater than or equal to threshold value TH1. Figure 4 also illustrates that the target category to which the advertising target of primary advertisement A belongs is category CT1, and the target category to which the advertising target of secondary advertisement B belongs is category CT2. Figure 4 also illustrates an example of a composite advertisement delivery result in which the click-through rate, which indicates the advertising effectiveness of a composite advertisement consisting of primary advertisement C and secondary advertisement D, is less than threshold value TH1. Figure 4 also illustrates that the target category to which the advertising target of primary advertisement C belongs is category CT3, and the target category to which the advertising target of secondary advertisement D belongs is category CT4.

The information processing device 100 selects a primary advertisement and a secondary advertisement, avoiding combinations of advertisements that constitute a combined advertisement for which the index value indicating the advertising effectiveness of the combined advertisement does not meet a predetermined standard. Specifically, the information processing device 100 first identifies target categories for selecting the primary advertisement and secondary advertisement to be combined. For example, in the case shown in FIG. 4, the information processing device 100 identifies category CT1, to which the advertising target of advertisement A that constitutes a combined advertisement and whose click-through rate is equal to or greater than threshold TH1, belongs, as the target category for selecting the primary advertisement, and similarly identifies category CT2, to which the advertising target of advertisement B belongs, as the target category for selecting the secondary advertisement.

Next, the information processing device 100 acquires primary and secondary advertisements that satisfy predetermined conditions from among multiple advertisements belonging to the target category. For example, the information processing device 100 may select a first advertisement and a second advertisement based on the amount of advertising expenses associated with each advertisement. In the case shown in FIG. 4, the information processing device 100 randomly selects, as primary advertisements, from multiple advertisements whose advertising targets belong to category CT1, advertisements that satisfy the condition that the amount of advertising expenses is equal to or greater than threshold TH2. In the case shown in FIG. 4, the information processing device 100 randomly selects, as secondary advertisements, from multiple advertisements whose advertising targets belong to category CT2, advertisements that satisfy the condition that the amount of advertising expenses is less than threshold TH2.

In addition, the information processing device 100 may select a primary advertisement and a secondary advertisement based on information indicating the degree to which the advertiser of each advertisement is recognized (hereinafter referred to as "recognition"). In the case shown in FIG. 4, the information processing device 100 randomly selects, as a primary advertisement, an advertisement that satisfies the condition that its recognition is equal to or greater than a threshold value TH3 from among multiple advertisements whose advertising target belongs to category CT1. In the case shown in FIG. 4, the information processing device 100 randomly selects, as a secondary advertisement, an advertisement that satisfies the condition that its recognition is less than a threshold value TH3 from among multiple advertisements whose advertising target belongs to category CT2.

Then, the information processing device 100 generates a composite advertisement using the primary advertisement and the secondary advertisement selected in step S41 (step S42). For example, the information processing device 100 generates a composite advertisement by using the primary advertisement and the secondary advertisement selected as the objects to be combined to combine the secondary advertisement with a portion of the primary advertisement. Figure 4 shows an example of a composite advertisement AD_a in which advertisement B selected as a secondary advertisement is combined with a portion of advertisement A selected as the primary advertisement.

In addition, the information processing device 100 may select the primary and secondary advertisements by taking into account information about combinations of advertisements that should be avoided. This allows the information processing device 100 to avoid inappropriate advertising language when delivering a composite advertisement.

In addition, the information processing device 100 may select the primary and secondary advertisements by taking into account information indicating the competitive relationships between advertisers. This allows the information processing device 100 to take care, when delivering a composite advertisement, to ensure that the market share of the transaction subject, such as a product or service offered by the advertiser of the primary advertisement, is not eroded.

(1-5. Advanced synthetic ad delivery methods)
Furthermore, the information processing device 100 may select advertisements for conveying information about advertisers different from the other advertisers from among advertisements submitted by the other advertisers, and distribute a composite advertisement generated by combining the selected advertisements. This allows the information processing device 100 to realize novel advertisement distribution without being bound by existing methods of advertisement distribution. Below, an overview of an advanced version of the composite advertisement distribution method according to the embodiment will be described with reference to FIG. 5. FIG. 5 is a diagram showing an overview of an advanced version of the composite advertisement distribution method according to the embodiment.

As shown in FIG. 5, advertiser U2_D uses advertiser terminal 20 to send a composite advertisement generation request to information processing device 100 (step S51). FIG. 5 shows an example in which the industry of the business run by advertiser U2_D is "industry X4." Note that the industry can be set at any granularity.

The advertiser terminal 20 shown in FIG. 5 is an information processing terminal used by advertiser U2_D, a user of various services provided by information processing device 100. Possible advertiser terminals 20 include smartphones, tablet terminals, notebook PCs (Personal Computers), desktop PCs, mobile phones, and PDAs (Personal Digital Assistants). In the example of FIG. 5, a smartphone is shown as the advertiser terminal 20. Note that in the following description, the advertiser terminal 20 may be referred to as advertiser U2_D. In other words, advertiser U2_D can be read as advertiser terminal 20.

The advertiser terminal 20 also has a communication unit for connecting to the network N (see FIG. 6) via a wireless communication network such as LTE (Long Term Evolution), 4G (4th Generation), or 5G (5th Generation: fifth generation mobile communication system), or via short-range wireless communication such as Bluetooth (registered trademark) or wireless LAN (Local Area Network). The advertiser U2_D operates the advertiser terminal 20 having the communication unit to access the information processing device 100 and use various services.

The advertiser terminal 20 can also display information for using various services provided by the information processing device 100 using a web browser or application. When the advertiser terminal 20 receives control information from the information processing device 100 for displaying information using a web browser or application, the advertiser terminal 20 performs the display processing in accordance with the received control information.

Meanwhile, when the information processing device 100 receives a composite advertisement generation request from the advertiser U2_D, it selects multiple advertisements from among advertisements submitted by advertisers other than the requesting advertiser U2_D as materials for generating a composite advertisement to convey information about the requesting advertiser U2_D requesting the generation of the composite advertisement (step S52). For example, the information processing device 100 identifies related industries that are associated with the industry corresponding to the requesting advertiser requesting the generation of the composite advertisement based on a predetermined industry correspondence relationship. Figure 5 shows industry correspondence information indicating the industry correspondence relationship. According to the industry correspondence information shown in Figure 5, an example is shown in which the industry of the business operated by the advertiser U2_A is "industry X1," the industry of the business operated by the advertiser U2_B is "industry X2," the industry of the business operated by the advertiser U2_C is "industry X3," and the industry of the business operated by the advertiser U2_D is "industry X4." In the case shown in FIG. 5, the information processing device 100 references industry correspondence information that indicates the correspondence between predefined industries, and identifies "industry X1, industry X2" as related industries associated with the industry "industry X4" that corresponds to the requesting advertiser U2_D that is requesting the generation of a composite advertisement.

The information processing device 100 then selects advertisements to be combined from among advertisements entered by advertisers associated with the identified industry. Figure 5 shows an example in which advertisements have been submitted by advertisers U2_A, U2_B, and U2_C, who are different from advertiser U2_D, the requester requesting the generation of the combined advertisement. Note that the advertisements may also have been submitted by advertiser U2_D, who sent the request to generate the combined advertisement.

In the example shown in FIG. 5, the information processing device 100 selects advertisements to be combined from advertisements submitted by advertiser U2_A linked to "Industry X1," identified as a related industry of "Industry X4," the industry corresponding to advertiser U2_D, and advertisements submitted by advertiser U2_B linked to "Industry X2," identified as a related industry of "Industry X4," the industry corresponding to advertiser U2_D. In the example shown in FIG. 5, the information processing device 100 has identified multiple related industries (e.g., industry X1 and industry X2) associated with the industry corresponding to advertiser U2_D, the requester requesting the generation of the combined advertisement. Therefore, the information processing device 100 selects advertisements to be combined from each of the advertisements submitted by advertisers linked to each of the identified industries. That is, the information processing device 100 selects advertisements to be combined from advertisements submitted by advertiser U2_A associated with "industry X1," and also selects advertisements to be combined from advertisements submitted by advertiser U2_B associated with "industry X2." FIG. 5 shows an example in which "Advertisement A-1" is selected from advertisements submitted by advertiser U2_A, and "Advertisement B-1" is selected from advertisements submitted by advertiser U2_B. Note that if only one related industry is identified that is associated with the industry corresponding to advertiser U2_D, the requester requesting the generation of a combined advertisement, multiple advertisements to be combined may be selected from advertisements submitted by advertisers associated with the identified industry.

The information processing device 100 also generates a composite advertisement by combining the advertisements selected in step S52 (step S53). Figure 5 shows an example of a composite advertisement AD_b that combines "Advertisement A-1" and "Advertisement B-1." The information processing device 100 then delivers the composite advertisement generated in step S53 to user U1 (step S54).

Furthermore, when the advertising materials of each advertisement to be combined are images, the information processing device 100 may generate the combined advertisement using a GAN (Generative Adversarial Network).

Furthermore, if advertiser U2_D, which sent the composite advertisement generation request, achieves a predetermined result, the information processing device 100 may return a reward according to the predetermined result to advertiser U2_A of "Advertisement A-1" and advertiser U2_B of "Advertisement B-1," which were selected as advertisements for conveying information about advertiser U2_D.

In addition, the information processing device 100 may change the advertisement to be selected depending on the user attributes (context) to which the composite advertisement is to be delivered.In addition, the information processing device 100 may change the landing page to be delivered to the user depending on the user attributes (context) of the user who selected (clicked) the composite advertisement.

Note that Figure 5 describes an example of information processing in which the information processing device 100 identifies a related industry associated with the industry corresponding to the requesting advertiser requesting the generation of a composite advertisement based on the correspondence relationship with the advertiser's industry, and selects advertisements to be combined from advertisements entered by advertisers linked to the identified industry. However, this does not need to be particularly limited. Industry is an example of information that can identify the correspondence relationship between advertisers, and processing can be performed using any information other than industry as long as it can identify the correspondence relationship between advertisers. For example, the information processing device 100 may perform processing based on categories arbitrarily set for each advertiser. Specifically, the information processing device 100 can identify related categories associated with categories corresponding to the requesting advertiser requesting the generation of a composite advertisement based on the category correspondence relationship previously set for each advertiser, and select advertisements to be combined from advertisements entered by advertisers linked to the identified related categories. Furthermore, when multiple related categories are identified, the information processing device 100 can select advertisements to be combined from each of the advertisements submitted by advertisers linked to each of the identified related categories.

2. Configuration of information processing device
An information processing device 100 according to an embodiment will be described with reference to Fig. 6. Fig. 6 is a diagram showing an example of the configuration of the information processing device according to an embodiment. As shown in Fig. 6, the information processing device 100 includes a communication unit 110, a storage unit 120, and a control unit 130.

(Regarding the communication unit 110)
The communication unit 110 is realized by, for example, a network interface card (NIC). The communication unit 110 is connected to the network N by wire or wirelessly. The information processing device 100 transmits and receives information to and from the user terminal 10 and the advertiser terminal 20 via the network N.

(Regarding the storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 120 has an advertisement information storage unit 121, a distribution result information storage unit 122, an advertising effectiveness estimation model storage unit 123, an advertising vector output model storage unit 124, and an industry correspondence information storage unit 125.

In addition to the information shown in FIG. 6, the memory unit 120 can store information to be referenced when selecting multiple advertisements to be combined, such as information regarding combinations of advertisements that should be avoided and information indicating competitive relationships between advertisers.In addition to the information shown in FIG. 6, the memory unit 120 can store selection rules for automatically selecting a primary advertisement and a secondary advertisement from multiple advertisements to be combined.

(Advertisement information storage unit 121)
The advertisement information storage unit 121 stores advertisement information related to advertisements. Fig. 7 is a diagram showing an example of advertisement information according to an embodiment. Note that Fig. 7 shows an example of advertisement information according to an embodiment, and the advertisement information may include information other than that shown in Fig. 7. Also, Fig. 7 shows an example in which conceptual information is displayed for multiple items included in the advertisement information, but in reality, specific information such as numerical values and character strings is stored.

As shown in Figure 7, the advertising information stored in the advertising information storage unit 121 has multiple items such as an "advertising ID," an "advertiser ID," an "advertising detail information," a "category information," a "related keyword information," an "advertising type," an "advertising budget," "awareness," and an "industry." These items in the advertising information are associated with each other.

The "Advertisement ID" field stores identification information (advertisement ID) for identifying an advertisement. Note that even in the case of a composite advertisement, the "Advertisement ID" field may store a single ad ID assigned to the composite advertisement. Furthermore, in the case of a composite advertisement, the "Advertisement ID" field may also store ad IDs assigned to each advertisement that makes up the composite advertisement.

The "Advertiser ID" field stores identification information (advertiser ID) for identifying the advertiser who submitted the advertisement. In the case of a composite advertisement, the "Advertiser ID" field may also store the advertiser IDs of each advertisement that makes up the combined advertisement.

The "Detailed Advertisement Information" field stores detailed information about the content of the advertisement. Examples of detailed information include the title displayed as text related to the advertisement, a description summarizing the content of the advertising page (web page) on which the advertisement is displayed, and the URL (Uniform Resource Locator) of the link destination set in the displayed advertisement.

The "Category Information" field stores information indicating the target category to which the advertising target belongs. Target categories are indicated by classifications that group identical or similar products or identical or similar services at a specified granularity.

The "Related Keyword Information" field stores information indicating related keywords associated with the advertisement. For example, for an advertisement about automobiles, related keywords would be words that tend to co-occur in search queries, such as "new car" and "sales."

The "Ad Type" field stores information indicating the type of advertisement set by the advertiser when submitting the advertisement. Ad types include three types: "Normal," "Parasitic Allowed," and "Parasitic Desired." "Normal" is a type selected by advertisers who wish to be delivered as a conventional advertisement rather than as a composite advertisement. "Parasitic Allowed" is a type selected by advertisers who wish to allow advertisements from other advertisers to be combined as a composite advertisement. Advertisements with "Parasitic Allowed" set are treated as primary advertisements. "Parasitic Desired" is a type selected by advertisers who wish to be combined with advertisements from other advertisers as a composite advertisement. Advertisements with "Parasitic Desired" set are treated as secondary advertisements.

The "Advertising Budget" field stores information indicating the advertising budget (advertising expenses) set by the advertiser when submitting the advertisement.

The "Awareness" field stores information indicating the degree to which the advertiser is recognized by users (awareness). The awareness stored in the "Awareness" field may be predetermined based on, for example, the number of followers on a social networking service (SNS), the number of searches on a search service, or information collected by any method, such as user surveys or crowdsourcing, regarding each advertiser.

The "Industry" field stores information indicating the type of business run by the advertiser. The industry is indicated by a classification that compiles the business run by the advertiser at a specified level of granularity.

For example, according to Figure 7, the advertiser of the advertising information identified by the advertising ID: "Advertisement #01" is identified by "Advertiser #01." Also, according to Figure 7, information mutually associated with the advertising information identified by the advertising ID: "Advertisement #01" includes "advertising identification information EX01," "category information #01," "related keyword information #01," "advertising type EX01," "advertising budget EX01," "recognition EX01," and "industry EX01."

(Distribution result information storage unit 122)
The distribution result information storage unit 122 stores distribution result information relating to the distribution results of the advertisement. Fig. 8 is a diagram illustrating an example of the distribution result information according to the embodiment.

Note that Figure 8 shows an example of delivery result information according to an embodiment, and the delivery result information may include information other than that shown in Figure 8. For example, the delivery result information storage unit 122 may store identification information (advertiser ID) for identifying the advertiser, and predetermined index values indicating advertising effectiveness such as cost per click (CPC), click-through rate (CTR), number of conversions, and page views (PV).

In addition, Figure 8 shows an example in which conceptual information is displayed for multiple items in the distribution result information, but in reality, specific information such as numerical values or character strings corresponding to each item is stored.

As shown in FIG. 8, the distribution result information stored in the distribution result information storage unit 122 has multiple items such as an "advertising ID," an "impression count," a "click count," and a "conversion rate." These items in the distribution result information are associated with each other.

The "Advertising ID" field stores identification information (advertising ID) for identifying the advertisement. Note that the identification information (advertising ID) stored in the "Advertising ID" field of the distribution result information may be the same as the identification information (advertising ID) stored in the "Advertising ID" field of the advertising information shown in Figure 7.

The "Number of impressions" field stores information indicating the number of times an advertisement was displayed (number of impressions). The "Number of clicks" field stores information indicating the number of times an advertisement was clicked. The "Conversion rate" field stores information indicating the percentage of users who accessed a specified web page and achieved a specified result (conversion).

For example, Figure 8 shows that the number of times an advertisement identified by the advertisement ID: "Advertisement #01" has been displayed is "Number of Display Times EX01", the number of clicks is "Number of Clicks EX01", and the conversion rate is "Conversion Rate EX01".

(Advertising effectiveness estimation model storage unit 123)
The advertising effectiveness estimation model memory unit 123 stores information about an advertising effectiveness estimation model, which is a trained model that inputs information about the primary advertisement and the secondary advertisement to be combined and outputs an estimated value of the advertising effectiveness obtained by the combined advertisement obtained by combining them.

The information about the advertising effectiveness estimation model stored in the advertising effectiveness estimation model storage unit 123 includes various pieces of information that make up the model, such as information about the model's configuration (network configuration) and information about parameters. In other words, the various pieces of information that make up the model include information about the nodes in each layer of the network, the functions used by each node, the connection relationships between the nodes, and the connection coefficients set for the connections between the nodes.

(Advertising Vector Output Model Storage Unit 124)
The advertising vector output model storage unit 124 inputs information about the primary advertisement and the secondary advertisement to be synthesized, and stores information about the advertising vector output model, which is a trained model that outputs a feature vector obtained by vectorizing the features of the information about the primary advertisement and a feature vector obtained by vectorizing the features of the information about the secondary advertisement.

The information about the advertising vector output model stored in the advertising vector output model storage unit 124 includes various pieces of information that make up the model, such as information about the model's configuration (network configuration) and information about parameters. In other words, the various pieces of information that make up the model include information about the nodes in each layer of the network, the functions used by each node, the connection relationships between the nodes, and the connection coefficients set for the connections between the nodes.

(Industry Correspondence Information Storage Unit 125)
The business type correspondence information storage unit 125 stores information indicating a correspondence relationship between predefined business types and information on related business types related to the business type of the business run by the advertiser. Fig. 9 is a diagram illustrating an example of the business type correspondence information according to the embodiment.

As shown in Figure 9, the industry correspondence information stored in the industry correspondence information storage unit 125 has multiple items, such as an "industry" item and an "associated industry" item. These items in the industry correspondence information are associated with each other.

The "Industry" field stores information indicating the industry that represents the business conducted by the advertiser, grouped into categories at a predetermined level of granularity. The "Related Industry" field stores information indicating related industries that have a certain degree of relevance to the industry stored in the "Industry" field.

For example, Figure 9 shows that the related industries of "Industry 1" are "Industry 4, Industry 5." When the information processing device 100 receives a request to generate a composite advertisement from an advertiser, it can identify the related industries related to the industry corresponding to the advertiser by referring to the industry correspondence information.

(Regarding the control unit 130)
The control unit 130 is realized by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like executing various programs stored in a storage device inside the information processing device 100 using RAM as a work area. The control unit 130 is also realized by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

As shown in FIG. 6, the control unit 130 has a selection unit 131, a generation unit 132, a distribution unit 133, a determination unit 134, and a learning unit 135, and realizes or executes the information processing functions and actions described below. Note that the internal configuration of the control unit 130 is not limited to the configuration shown in FIG. 6, and may be other configurations that perform the information processing described below. Furthermore, the connection relationships between the processing units in the control unit 130 are not limited to the connection relationships shown in FIG. 6, and may be other connection relationships.

(Selection unit 131)
The selection unit 131 selects a primary advertisement and a secondary advertisement. Furthermore, the selection unit 131 can select a primary advertisement and a secondary advertisement from among a plurality of advertisements to be combined, based on information about advertisement types preset for each advertisement by the advertiser. Furthermore, the selection unit 131 can select a primary advertisement and a secondary advertisement while avoiding a combination of advertisements for which an index value indicating the advertising effectiveness of the combined advertisement is less than a predetermined standard. Furthermore, the selection unit 131 can select a primary advertisement and a secondary advertisement while prioritizing a combination of advertisements for which an index value indicating the advertising effectiveness of the combined advertisement is equal to or greater than a predetermined standard. Furthermore, the selection unit 131 can select a primary advertisement and a secondary advertisement while taking into account information about a combination of advertisements that should be avoided. Furthermore, the selection unit 131 can select a primary advertisement and a secondary advertisement while taking into account information indicating a competitive relationship between advertisers.

The selection unit 131 can also select a primary advertisement and a secondary advertisement based on the output of a trained model trained by the learning unit 135. For example, the selection unit 131 can select a primary advertisement and a secondary advertisement to be combined based on the output from an advertising effectiveness estimation model, which is a trained model trained by the learning unit 135. For example, the selection unit 131 can also select a primary advertisement and a secondary advertisement to be combined based on the similarity between feature vectors output from an advertising vector output model, which is a trained model trained by the learning unit 135.

The selection unit 131 can also select a primary advertisement and a secondary advertisement from among multiple advertisements to be combined, in accordance with predefined selection rules, based on the delivery results of the combined advertisement. For example, the selection unit 131 can select a primary advertisement and a secondary advertisement while avoiding combinations of advertisements that make up a combined advertisement for which the index value indicating the advertising effectiveness of the combined advertisement does not satisfy a predetermined standard. For example, the selection unit 131 can also select a primary advertisement and a secondary advertisement based on the amount of advertising expenses associated with each advertisement. For example, the selection unit 131 can also select a primary advertisement and a secondary advertisement based on information indicating the degree to which the advertiser of each advertisement is recognized.

The selection unit 131 can also select multiple advertisements from advertisements submitted by advertisers other than the requesting advertiser as material for generating a composite advertisement for conveying information about the requesting advertiser requesting the generation of a composite advertisement. For example, the selection unit 131 can identify related categories associated with the category corresponding to the requesting advertiser based on the correspondence between categories (e.g., industry) pre-set for each advertiser, and select advertisements to be combined from advertisements submitted by advertisers linked to the identified related categories. For example, when multiple related categories are identified, the selection unit 131 can select advertisements to be combined from each of the advertisements submitted by advertisers linked to each of the identified related categories.

(Generation unit 132)
The generation unit 132 generates a composite advertisement by combining a part of the primary advertisement with the secondary advertisement, based on the primary advertisement and the secondary advertisement selected by the selection unit 131. The generation unit 132 can arbitrarily determine the layout and display size of the secondary advertisement relative to the primary advertisement, provided that at least the advertising target of the primary advertisement is recognizable to the user.

(Distribution unit 133)
The distribution unit 133 distributes the combined advertisement generated by the generation unit 132 via the communication unit 110. The distribution unit 133 collects distribution results of the combined advertisement and registers distribution result information related to the collected distribution results in the distribution result information storage unit 122.

(Determination unit 134)
When a composite advertisement is selected, the determination unit 134 determines the amount to be charged to each advertiser associated with the composite advertisement. For example, the determination unit 134 may charge only the advertiser of the primary advertisement. In this case, the determination unit 134 may set the amount to be charged to the advertiser of the primary advertisement lower than the cost per click corresponding to a "normal" advertisement. Furthermore, the determination unit 134 may charge the advertiser corresponding to the clicked advertisement of the primary advertisement and the secondary advertisements that constitute the composite advertisement. In this case, the determination unit 134 may prorate the sum of the cost per click of the primary advertisement and the cost per click corresponding to the secondary advertisement according to the display size of the composite advertisement, and determine the prorated amount as the amount to be charged.

(Learning unit 135)
The learning unit 135 causes the model to learn characteristics of compatibility when combining a primary advertisement and a secondary advertisement based on the delivery result of the combined advertisement. For example, the learning unit 135 causes the model to learn a correspondence between information about the primary advertisement and the secondary advertisement, and an approval result for acceptance of the secondary advertisement by the advertiser of the primary advertisement and a score based on the advertising effectiveness of the combined advertisement. As a result, the learning unit 135 inputs information about the primary advertisement and the secondary advertisement to be combined, and generates an advertising effectiveness estimation model, which is a trained model that outputs an estimated value of advertising effectiveness obtained by a combined advertisement obtained by combining the information about the primary advertisement and the secondary advertisement.

The information about the primary and secondary advertisements may be any information related to the advertisement. For example, the information about the advertisement may be a title displayed as text related to the advertisement, a description summarizing the content of the advertising page (web page) on which the advertisement is displayed, or information about the content of the advertisement, such as images or videos used as advertising materials. Furthermore, if the advertiser is a company, the information about the advertisement may be information such as capital and number of employees, or if the advertiser is a listed company, it may be stock price information.

The approval result for the primary advertiser's acceptance of the secondary advertisement is a history of information indicating whether the primary advertiser approved or denied the acceptance of the secondary advertisement as part of the primary advertisement. The advertising effectiveness of the composite advertisement is determined based on a predetermined indicator of advertising effectiveness. The predetermined indicator of advertising effectiveness can be any indicator, such as cost per acquisition (CPA), cost per click (CPC), click-through rate (CTR), number of conversions, conversion rate (CVR), or page view (PV).

For example, the administrator of the information processing device 100 pre-sets a score that reflects the result of approval of the acceptance of the secondary advertisement by the advertiser of the primary advertisement and an index value that indicates the advertising effectiveness of the combined advertisement. Specifically, the administrator sets the score so that the score increases as the acceptance of the secondary advertisement is approved and the index value increases, and decreases when the acceptance of the secondary advertisement is rejected. The administrator then associates the set score with a combination of information about the primary advertisement and the secondary advertisement, and creates a learning data sample.

The learning unit 135 uses the created learning data samples to train a model to understand the correspondence between information about the primary and secondary advertisements, the approval result for the acceptance of the secondary advertisement by the advertiser of the primary advertisement, and a score based on the advertising effectiveness of the composite advertisement. The learning unit 135 can use any network, such as a neural network, as a learning model to train the model to understand the correspondence. Furthermore, when a neural network is used as the learning model, the learning unit 135 performs a learning process to adjust the values of the weights (connection coefficients) that are taken into account when scores (values) are propagated between nodes in each layer that make up the neural network. For example, the learning unit 135 uses any method, such as backpropagation, to optimize (correct) the parameters (connection coefficients) so as to reduce the error between the output from the learning model and the correct answer (correct answer data) corresponding to the input. When using the backpropagation method, the learning unit 135 executes the learning process to minimize a predetermined loss function, thereby correcting the parameters (connection coefficients) to reduce the error between the output from the learning model and the correct answer (correct answer data) corresponding to the input.

Through the above-described learning process, the learning unit 135 can generate an advertising effectiveness estimation model, which is a trained model that receives information about the primary and secondary advertisements to be combined and outputs an estimated value of the advertising effectiveness obtained from a combined advertisement obtained by combining these advertisements. The learning unit 135 registers information about the generated advertising effectiveness estimation model in the advertising effectiveness estimation model storage unit 123.

The learning unit 135 also learns a model that vectorizes each advertisement based on information about the advertisers of each advertisement that makes up the composite advertisement and the delivery results of the composite advertisement. As a result, the learning unit 135 inputs information about the primary advertisement and secondary advertisements to be combined, and generates an advertisement vector output model, which is a trained model that outputs a feature vector obtained by vectorizing the features of the information about the primary advertisement and a feature vector obtained by vectorizing the features of the information about the secondary advertisement.

For example, the learning unit 135 can train a model so that the closer the scale of the advertisers, the smaller the similarity between the vectors corresponding to each advertisement that makes up the composite advertisement. Furthermore, the learning unit 135 can train a model so that the more frequently the composite advertisement is selected, the smaller the similarity between the vectors corresponding to each advertisement that makes up the composite advertisement.

The administrator of the information processing device 100 sets information about the advertisers of the primary and secondary advertisements, the approval result for the primary advertisement's acceptance of the secondary advertisement, and a score based on the advertising effectiveness of the combined advertisement.

The information about the advertisers of the primary and secondary advertisements may be any information related to the advertisement. For example, if the advertiser is a company, it may be information such as capital and number of employees, and if it is a listed company, it may be information about the stock price. Note that information about the advertisers of the primary and secondary advertisements may also be used when setting the score. In this case, the information about the advertisement may be information about the content of the advertisement, such as the title displayed as text related to the advertisement, a description summarizing the content of the advertising page (web page) on which the advertisement is displayed, or images and videos used as advertising materials.

The approval result for the primary advertiser's acceptance of the secondary advertisement is a history of information indicating whether the primary advertiser approved or denied the acceptance of the secondary advertisement as part of the primary advertisement. The advertising effectiveness of the composite advertisement is determined based on a predetermined indicator of advertising effectiveness. The predetermined indicator of advertising effectiveness can be any indicator, such as cost per acquisition (CPA), cost per click (CPC), click-through rate (CTR), number of conversions, conversion rate (CVR), or page view (PV).

For example, the administrator of the information processing device 100 pre-sets a score that reflects the result of approval of the acceptance of the secondary advertisement by the advertiser of the primary advertisement and an index value that indicates the advertising effectiveness of the combined advertisement. Specifically, the administrator sets the score so that the score increases as the acceptance of the secondary advertisement is approved and the index value increases, and decreases when the acceptance of the secondary advertisement is rejected. The administrator then associates the set score with a combination of information about the advertisers of the primary advertisement and the secondary advertisement, and creates a learning data sample.

The learning unit 135 uses the created training data sample to train a model so that the larger the score set for a combination of information about the advertisers of the primary advertisement and the secondary advertisement, the smaller the similarity between the feature vector corresponding to the advertiser of the primary advertisement and the feature vector corresponding to the advertiser of the secondary advertisement is output as a vector. Specifically, the learning unit 135 trains a model to output each feature vector so that the larger the set score in the training data sample, the smaller the cosine similarity between the feature vector corresponding to the primary advertisement and the feature vector corresponding to the secondary advertisement.

For example, when using an arbitrary network such as a neural network as a training model, the learning unit 135 performs a training process to adjust the values of weights (connection coefficients) that are taken into account when scores (values) are propagated between nodes in each layer that make up the neural network. For example, the learning unit 135 uses an arbitrary method such as backpropagation to perform a process to optimize (correct) parameters (connection coefficients) so as to reduce the error between the output from the training model and the correct answer (correct data) corresponding to the input. In other words, the learning unit 135 performs a process to optimize parameters (connection coefficients) so that the cosine similarity between the feature vector corresponding to the primary advertisement and the feature vector corresponding to the secondary advertisement decreases as the set score in the training data sample increases. Note that when using the backpropagation method, the learning unit 135 can perform a training process to minimize a predetermined loss function, thereby correcting the parameters (connection coefficients) so as to reduce the error between the output from the training model and the correct answer (correct data) corresponding to the input.

Through the above-described learning process, the learning unit 135 can generate an advertising vector output model, which is a trained model that vectorizes each advertisement, based on information about the advertisers of each advertisement that makes up the composite advertisement and the delivery results of the composite advertisement. The learning unit 135 registers information about the generated advertising vector output model in the advertising vector output model storage unit 124.

3. Processing Procedure
(3-1. Synthetic Ad Delivery Procedure (Part 1))
Hereinafter, an example of an information processing procedure executed by the information processing device 100 according to the embodiment will be described with reference to the drawings. Hereinafter, an example of a composite advertisement delivery procedure (part 1) according to the embodiment will be described with reference to FIG. 10. FIG. 10 is a flowchart showing an example of a composite advertisement delivery procedure (part 1) according to the embodiment. The processing procedure shown in FIG. 10 is executed by the control unit 130 of the information processing device 100. The processing procedure shown in FIG. 10 is repeatedly executed while the information processing device 100 is operating.

As shown in FIG. 10, the selection unit 131 selects a primary advertisement and a secondary advertisement from among multiple advertisements to be combined (step S101).

In addition, the generation unit 132 generates a composite advertisement by combining a portion of the primary advertisement with the secondary advertisement based on the primary advertisement and secondary advertisement selected by the selection unit 131 (step S102).

The distribution unit 133 also distributes the composite advertisement generated by the generation unit 132 via the communication unit 110 (step S103), and ends the processing procedure shown in Figure 10.

(3-2. Synthetic Ad Delivery Procedure (Part 2))
An example of a procedure (part 2) for delivering a composite advertisement according to the embodiment will be described below with reference to Fig. 11. Fig. 11 is a flowchart showing an example of a procedure (part 2) for delivering a composite advertisement according to the embodiment. The processing procedure shown in Fig. 11 is executed by the control unit 130 of the information processing device 100. The processing procedure shown in Fig. 11 is repeatedly executed while the information processing device 100 is operating.

As shown in FIG. 11, the selection unit 131 selects a primary advertisement and a secondary advertisement from among multiple advertisements to be combined, in accordance with predefined selection rules, based on the delivery results of the combined advertisement (step S201).

In addition, the generation unit 132 generates a combined advertisement by combining a portion of the primary advertisement with the secondary advertisement based on the primary advertisement and secondary advertisement selected by the selection unit 131 (step S202).

The distribution unit 133 also distributes the composite advertisement generated by the generation unit 132 via the communication unit 110 (step S203), and ends the processing procedure shown in Figure 11.

(3-3. Synthetic Ad Delivery Procedure (Part 3))
Hereinafter, an example of a procedure (part 3) for delivering a composite advertisement according to the embodiment will be described with reference to Fig. 12. Fig. 12 is a flowchart showing an example of a procedure (part 3) for delivering a composite advertisement according to the embodiment. The processing procedure shown in Fig. 12 is executed by the control unit 130 of the information processing device 100. The processing procedure shown in Fig. 12 is repeatedly executed while the information processing device 100 is operating.

As shown in FIG. 12, the selection unit 131 selects multiple advertisements from advertisements submitted by other advertisers to convey information about advertisers different from the other advertisers (step S301).

Next, the distribution unit 133 distributes the combined advertisement generated by combining the selected advertisements via the communication unit 110 (step S302), and ends the processing procedure shown in FIG. 12.

4. Modifications
The information processing device 100 according to the above embodiment may be implemented in various different forms other than the above embodiment, so other embodiments of the information processing device 100 will be described below.

(4-1. Display of composite advertisements)
In the above embodiment, the information processing device 100 may separately count user responses, such as the number of clicks on the composite advertisement, for the primary advertisement and the secondary advertisements that make up the composite advertisement, and dynamically change the display mode of the composite advertisement based on the counting results. For example, the information processing device 100 may dynamically change the display size of an image, the saturation of the display color of the image, the transparency of the image, etc., depending on the cumulative number of clicks. Specifically, the information processing device 100 may increase the display size, darken the display color, or reduce the transparency of the advertisement that has the greater cumulative number of clicks, out of the primary advertisement and the secondary advertisement.

Furthermore, in the above embodiment, the information processing device 100 may change the secondary advertisements that make up the composite advertisement to different advertisements in chronological order without changing the primary advertisement.

(4-2. Bidding for secondary advertisements)
In the above embodiment, the information processing device 100 may receive a selection of a primary advertisement desired to be parasitized from the advertiser of the secondary advertisement. Then, if there are overlapping requests, the information processing device 100 may execute a bidding process to win the parasitization rights.

(4-3. Remuneration to the main advertiser)
In the above embodiment, the information processing device 100 may kick back a predetermined reward to the advertiser of the main advertisement when the secondary advertisements that make up the composite advertisement achieve a predetermined result.

5. Hardware Configuration
The information processing device 100 according to the embodiment described above is realized, for example, by a computer 1000 configured as shown in Fig. 13. Fig. 13 is a hardware configuration diagram showing an example of a computer that realizes the functions of the information processing device according to the embodiment.

The computer 1000 is connected to an output device 1010 and an input device 1020, and has a configuration in which an arithmetic unit 1030, a primary storage device 1040, a secondary storage device 1050, an output IF (Interface) 1060, an input IF 1070, and a network IF 1080 are connected via a bus 1090.

The arithmetic device 1030 operates based on programs stored in the primary storage device 1040 and secondary storage device 1050, and programs read from the input device 1020, and executes various processes. The primary storage device 1040 is a memory device, such as RAM, that temporarily stores data used by the arithmetic device 1030 for various calculations. The secondary storage device 1050 is a storage device in which data used by the arithmetic device 1030 for various calculations and various databases are registered, and is realized by a ROM (Read Only Memory), HDD, flash memory, etc.

The output IF 1060 is an interface for transmitting information to be output to an output device 1010, such as a monitor or printer, which outputs various types of information. It is implemented, for example, by a connector conforming to a standard such as USB (Universal Serial Bus), DVI (Digital Visual Interface), or HDMI (High Definition Multimedia Interface). The input IF 1070 is an interface for receiving information from various input devices 1020, such as a mouse, keyboard, and scanner, and is implemented, for example, by USB.

The input device 1020 may be a device that reads information from optical recording media such as a CD (Compact Disc), DVD (Digital Versatile Disc), or PD (Phase Change Rewritable Disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. The input device 1020 may also be an external storage medium such as a USB memory.

Network IF 1080 receives data from other devices via network N and sends it to the computing device 1030, and also transmits data generated by the computing device 1030 to other devices via network N.

The arithmetic unit 1030 controls the output device 1010 and the input device 1020 via the output IF 1060 and the input IF 1070. For example, the arithmetic unit 1030 loads a program from the input device 1020 or the secondary storage device 1050 onto the primary storage device 1040 and executes the loaded program.

For example, when the computer 1000 functions as the information processing device 100 according to the above embodiment, the arithmetic unit 1030 of the computer 1000 executes a program (e.g., an information processing program) loaded onto the primary storage device 1040, thereby realizing functions similar to those of the control unit 130. In other words, the arithmetic unit 1030 works in cooperation with the program (e.g., an information processing program) loaded onto the primary storage device 1040 to realize the processing performed by the information processing device 100 according to the above embodiment.

[6. Other]
Of the processes described in the above embodiments, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically using a known method.In addition, the information including the processing procedures, specific names, various data and parameters shown in the above documents and drawings can be changed as desired unless otherwise specified.

Furthermore, the components of each device shown in the figure are functional concepts and do not necessarily have to be physically configured as shown. In other words, the specific form of distribution and integration of each device is not limited to that shown, and all or part of them can be functionally or physically distributed and integrated in any unit depending on various loads, usage conditions, etc.

Furthermore, the above-described embodiments can be combined as appropriate to the extent that the processing content is not contradictory.

7. Effects
The information processing device 100 according to the embodiment includes a selection unit 131 and a distribution unit 133. The selection unit 131 selects a plurality of advertisements from advertisements submitted by advertisers other than the requesting advertiser as materials for generating a composite advertisement for transmitting information about the requesting advertiser requesting generation of the composite advertisement. The distribution unit 133 distributes the composite advertisement generated by combining the advertisements selected by the selection unit 131.

In addition, in the information processing device 100 according to the embodiment, the selection unit 131 identifies related categories associated with the category corresponding to the requesting advertiser based on the correspondence between categories (for example, industry) previously set for each advertiser, and selects the advertisements to be combined from among advertisements input by advertisers that are linked to the identified related categories.

Furthermore, in the information processing device 100 according to the embodiment, when multiple related categories are identified, the selection unit 131 selects advertisements to be combined from each of the advertisements submitted by advertisers that are linked to each of the identified related categories.

For this reason, the information processing device 100 according to the embodiment can realize new ad delivery without being bound by existing methods of ad delivery, by using the processes executed by the above-mentioned units or any combination of the units.

The above describes in detail the embodiments of the present application based on several drawings, but these are merely examples, and the present invention can be implemented in other forms that incorporate various modifications and improvements based on the knowledge of those skilled in the art, including the aspects described in the Disclosure of the Invention section.

In addition, the above-mentioned "section, module, unit" can be interpreted as "means" or "circuit." For example, a control unit can be interpreted as a control means or a control circuit.

REFERENCE SIGNS LIST 10 User terminal 20 Advertiser terminal 100 Information processing device 110 Communication unit 120 Storage unit 121 Advertising information storage unit 122 Distribution result information storage unit 123 Advertising effectiveness estimation model storage unit 124 Advertising vector output model storage unit 125 Industry correspondence information storage unit 130 Control unit 131 Selection unit 132 Generation unit 133 Distribution unit 134 Determination unit 135 Learning unit

Claims (4)

a selection unit that selects a plurality of advertisements from advertisements submitted by advertisers other than the requesting advertiser as materials for generating a composite advertisement for transmitting information about the requesting advertiser requesting generation of the composite advertisement;
a distribution unit that distributes a composite advertisement generated by combining the advertisements selected by the selection unit,
The selection unit
An information processing device characterized by identifying a related industry that is associated with the industry corresponding to the requesting advertiser based on the correspondence relationship between the industries of the businesses run by the advertisers, and selecting the advertisements to be combined from advertisements entered by advertisers linked to the identified related industry . The selection unit
The information processing device according to claim 1 , wherein, when a plurality of related industries are identified, the advertisements to be combined are selected from advertisements submitted by advertisers associated with each of the identified related industries . 1. A computer-implemented information processing method, comprising:
a selection step of selecting a plurality of advertisements from advertisements submitted by advertisers other than the requesting advertiser as materials for generating a composite advertisement for transmitting information about the requesting advertiser requesting the generation of the composite advertisement;
a distribution step of distributing a composite advertisement generated by combining the advertisements selected in the selection step ,
The selection step includes:
Based on the correspondence relationship between the types of business conducted by the advertisers, a related business type associated with the business type corresponding to the requesting advertiser is identified, and the advertisement to be combined is selected from advertisements input by advertisers linked to the identified related business type.
1. An information processing method comprising: On the computer,
a selection step of selecting a plurality of advertisements from advertisements submitted by advertisers other than the requesting advertiser as materials for generating a composite advertisement for transmitting information about the requesting advertiser requesting the generation of the composite advertisement;
a distribution step of distributing a composite advertisement generated by combining the advertisements selected in the selection step ;
The selection procedure comprises:
An information processing program characterized by identifying a related industry that is associated with the industry corresponding to the requesting advertiser based on the correspondence relationship between the industries of the businesses run by the advertisers, and selecting the advertisements to be combined from advertisements entered by advertisers linked to the identified related industry .