Information processing device, information processing method, and information processing program

Abstract

To provide an information processing device capable of supporting the generation of information for content classification, an information processing method, and an information processing program.SOLUTION: An information processing device comprises an acquisition unit and a generation unit. The acquisition unit acquires classification condition information indicative of each of a plurality of classification conditions for classifying users. The generation unit generates a first content according to the classification conditions and used to classify a second content accessed by users on the basis of the classification condition information acquired by the acquisition unit.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to an information processing device, an information processing method, and an information processing program. [Background technology]

[0002] Conventionally, there is known a technique for classifying users based on information obtained from the users through a questionnaire. However, conducting a questionnaire is costly, and it is difficult to conduct a questionnaire every time a desired user classification is performed.

[0003] Therefore, the present applicant has proposed a technology for accepting characteristic information associated with a cluster, classifying content based on the characteristic information, and classifying users based on user access information to the classified content (see Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2017-21469 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, the above-mentioned technology has room for improvement in that feature information for classifying content needs to be extracted and generated from text written on the site, etc.

[0006] The present application has been made in consideration of the above, and has an object to provide an information processing device, an information processing method, and an information processing program that can assist in generating information for classifying content. [Means for solving the problem]

[0007] The information processing device according to the present application includes an acquisition unit and a generation unit. The acquisition unit acquires classification condition information indicating each of a plurality of classification conditions for classifying users. The generation unit generates, for each classification condition, a first content according to the classification condition and used to classify a second content accessed by the user, based on the classification condition information acquired by the acquisition unit. Effect of the Invention

[0008] Advantageous Effects of Invention According to one aspect of the embodiment, it is possible to provide an effect of assisting in the generation of information for classifying content. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a diagram for explaining information processing according to the embodiment. [Diagram 2] FIG. 2 is a diagram illustrating an example of the configuration of an information processing system according to the embodiment. [Diagram 3] FIG. 3 is a diagram illustrating an example of the configuration of the information processing device according to the embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of a user information table stored in the user information storage unit of the information processing device according to the embodiment. [Diagram 5] FIG. 5 is a diagram illustrating an example of a classification information table stored in the classification information storage unit of the information processing device according to the embodiment. [Figure 6] FIG. 6 is a diagram illustrating an example of a second content table stored in the first content storage unit of the information processing device according to the embodiment. [Figure 7] FIG. 7 is a diagram illustrating an example of a second content table stored in the second content storage unit of the information processing device according to the embodiment. [Figure 8] FIG. 8 is a diagram illustrating an example of a classification result table stored in the classification result storage unit of the information processing device according to the embodiment. [Figure 9] FIG. 9 is a diagram illustrating an example of input information input for generation by the generation unit of the information processing device according to the embodiment. [Figure 10] FIG. 10 is a diagram illustrating another example of input information input for generation by the generation unit of the information processing device according to the embodiment. [Figure 11] FIG. 11 is a diagram illustrating still another example of input information input for generation by the generation unit of the information processing device according to the embodiment. [Figure 12] FIG. 12 is a diagram illustrating another example of input information input for generation by the generation unit of the information processing device according to the embodiment. [Figure 13] FIG. 13 is a diagram illustrating yet another example of input information input for generation by the generation unit of the information processing device according to the embodiment. [Figure 14] FIG. 14 is a diagram showing an example of a classification request setting screen displayed on the terminal device based on the selection information transmitted by the providing unit of the information processing device according to the embodiment. [Figure 15] FIG. 15 is a flowchart illustrating an example of information processing by the processing unit of the information processing device according to the embodiment. [Figure 16] FIG. 16 is a flowchart illustrating an example of the classification process by the processing unit of the information processing device according to the embodiment. [Figure 17] FIG. 17 is a hardware configuration diagram showing an example of a computer that realizes the functions of the information processing device according to the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, the information processing device, the information processing method, and the information processing program according to the present application (hereinafter, referred to as "embodiments") will be described in detail with reference to the drawings. Note that the information processing device, the information processing method, and the information processing program according to the present application are not limited to these embodiments. In addition, each embodiment can be appropriately combined as long as the processing contents are not contradictory. In addition, the same parts in each of the following embodiments are given the same reference numerals, and duplicated explanations will be omitted.

[0011] [1. An example of information processing] First, an example of information processing according to the embodiment will be described with reference to Fig. 1. Fig. 1 is a diagram for explaining information processing according to the embodiment.

[0012] 1 is an information processing device that acquires various contents and access information from an information processing device 4 and classifies each user U who accessed the contents, and is realized by, for example, one or more servers or a cloud system. The access information is information on each user U's access to various contents.

[0013] The user U operates the terminal device 2 to use online sites provided by the information processing device 4. Online sites provided by the information processing device 4 include, but are not limited to, Q&A sites, news sites, EC (Electronic Commerce) sites, map sites, image posting sites, video viewing sites, etc. EC sites include, but are not limited to, shopping sites, auction sites, etc.

[0014] The information processing device 4 accepts posted content, which is content posted by each user U on the Q&A site, and provides the posted content to each user U. On the Q&A site, the posted content is question content, which is a question text, answer content, which is an answer text to the question text, etc.

[0015] Furthermore, the information processing device 4 accepts news articles as posted content at a news site, and provides the posted content to each user U. At a news site, the posted content includes, in addition to news articles, comments and ratings posted on the news articles.

[0016] Furthermore, the information processing device 4 accepts product content such as product descriptions that introduce products from sellers on the EC site as posted content, and provides such posted content to each user U. Furthermore, the posted content on the EC site includes, in addition to the product content, comments and ratings posted on the products.

[0017] Furthermore, the information processing device 4 accepts image content (e.g., still image content and video content) as posted content at an image posting site, and provides such posted content to each user U. Furthermore, the information processing device 4 accepts titles and caption texts of video content at a video viewing site, and provides such posted content to each user U.

[0018] Furthermore, the information processing device 4 accepts designation of a location or range from the user U at the map site, and provides map content showing a map of the location or range designated by the user U. The map content includes various information such as roads, various landmarks, various stores, routes, and stations.

[0019] Evaluations of news articles and product content can be positive or negative, and user U posts positive or negative evaluations by operating terminal device 2 to press a positive button or a negative button for the news article or product content.

[0020] 1, a worker O in charge of classifying a user U operates a terminal device 3 to input a plurality of classification conditions CC1 to CCn (n is an integer equal to or greater than 2) for classifying the user U and a target content type to the terminal device 3 (step S1). In the following, when the classification conditions CC1 to CCn are not individually indicated, they may be referred to as classification conditions CC.

[0021] The classification conditions CC are, for example, classification conditions based on the psychographic attributes of the user U, such as classification conditions based on the values ​​of the user U, classification conditions based on the preferences of the user U, classification conditions based on the lifestyle of the user U, etc., but are not limited to such examples.

[0022] In the example shown in Figure 1, the values ​​of user U are values ​​classified as freedom, equality, philanthropy, peace, etc., but are not limited to such examples and may be, for example, values ​​classified as traditionalism, success orientation, self-realization orientation, coexistence orientation, etc., or other values.

[0023] For example, when the object is a car, user U's values ​​may be classified as, for example, emphasis on safety, emphasis on livability, emphasis on design, emphasis on driving performance, etc., and when the object is a house, the values ​​may be classified as, for example, emphasis on location, emphasis on price, emphasis on layout, emphasis on design, emphasis on future potential, etc.

[0024] In addition, the values ​​of user U may be values ​​based on Schwartz's value theory. In this case, the values ​​of user U are values ​​classified as power, achievement, hedonism, stimulation, self-determination, universalism, philanthropy, tradition, harmony, and security, and the number of classification conditions CC is 10 (n=10).

[0025] The lifestyle of the user U may be classified, for example, by the AIO (Activities, Interests, Opinions) approach or VALS (Values ​​and LifeStyles). In the case of classification by the AIO approach, the classification conditions CC are indicated by classification conditions based on one or more items such as activities, interests, opinions, etc.

[0026] In the case of classification using VALS, the classification conditions CC are, for example, self-actualizers, successful people, group members, those who maintain their livelihood, those in poverty, those who want to succeed, those with social conscience, intellectuals, and young intellectuals, and the number of classification conditions CC is 9 (n=9).

[0027] The classification conditions CC may be, for example, classification conditions shown in the innovator theory that classifies the diffusion process of new products. In this case, the classification conditions CC are innovators, early adopters, early majority, late majority, and laggards, and the number of classification conditions CC is five (n=5).

[0028] The target content type is, for example, a type of the second content, and is indicated by, for example, a combination of an online site and a content type. For example, the target content type may be, but is not limited to, a news article on a news site, a comment on a news article on a news site, a product description on a product page on an EC site, a question on a Q&A site, an image content on an image posting site, a title and a caption text of a video content on a video viewing site, and the like.

[0029] The information processing device 1 can provide the terminal device 3 with selection information for selecting a classification method, a target content type, example sentences, etc. The selection information includes information on a plurality of classification methods, information on a plurality of target content types, and information on a plurality of example sentences, and each of the classification methods, target content types, and example sentences is displayed on the terminal device 3 in a selectable manner.

[0030] Examples of the multiple classification methods include, but are not limited to, a classification method based on multiple classification conditions included in past classification requests, a classification method based on values ​​based on Schwartz's value theory, a classification method based on the AIO approach, a classification method based on VALS, and a classification method based on classification conditions indicated in innovator theory.

[0031] By operating the terminal device 3, the worker O selects one classification method from among a plurality of classification methods, selects one or more target content types from among a plurality of target content types, and selects one or more example sentences from among a plurality of example sentences.

[0032] Next, the terminal device 3 transmits a classification request including classification condition information indicating each of the plurality of classification conditions CC1 to CCn input by the worker O in step S1 to the information processing device 1 (step S2).

[0033] In addition, when a classification method, target content type, example sentences, etc. are selected by the worker O, the terminal device 3 transmits a classification request to the information processing device 1, the classification request including information indicating each of the classification method, target content type, and example sentences selected by the worker O, as processing in step S2.

[0034] When the information processing device 1 receives a classification request transmitted from the terminal device 3, it acquires multiple pieces of classification condition information and information indicating the target content type contained in the classification request, and generates first contents CA1 to CAn (n is an integer greater than or equal to 2) according to the classification conditions CC1 to CCn based on the multiple pieces of classification condition information and the information indicating the target content type (step S3).

[0035] The first contents CA1 to CAn are contents according to the classification conditions CC1 to CCn and are contents of the target content type. For example, if the target content type is a question on a Q&A site, the first contents CA1 to CAn are questions likely to be posted on the Q&A site, and if the target content type is a news article on a news site, the first contents CA1 to CAn are news articles likely to be posted on the news site.

[0036] In the following, when the first contents CA1 to CAn are not individually distinguished from each other, they may be referred to as the first contents CA. The first contents CA are contents used to classify second contents to be accessed by the user U, which will be described later.

[0037] For example, if each classification condition CC indicated in the classification condition information is a value of a user U, the information processing device 1 generates content that is predicted to be accessed by a user U having the values ​​indicated in the classification condition CC as a first content CA for each classification condition CC.

[0038] In addition, when the classification request includes information indicating the classification method selected by the worker O, the information processing device 1 generates content predicted to be accessed by the user U who satisfies the classification condition CC corresponding to the classification method selected by the worker O as a first content CA for each classification condition CC.

[0039] The information processing device 1 generates the first content CA for each classification condition CC by using, for example, a generation AI (Artificial Intelligence). The generation AI is, for example, a text generation AI, an image generation AI, a multimodal AI, or the like.

[0040] The generation AI is arranged, for example, in an external information processing device, and the information processing device 1 acquires the first content CA for each classification condition CC by having the generation AI generate it via an API (Application Programming Interface) provided by the external information processing device, but is not limited to such an example. For example, the generation AI may be arranged in the information processing device 1.

[0041] A text generation AI is, for example, a language model trained to estimate and output the next token from an input token sequence, such as a transfer-based model or an RNN (Recurrent Neural Network)-based model.

[0042] Examples of the transfer-based model include, but are not limited to, GPT (Generative Pre-trained Transformer) and BARD (Bidirectional Auto Regressive Dialogues). Examples of the RNN-based model include, but are not limited to, RWKV (Receptance Weighted Key Value). It is preferable that the input information is trained so that it is not used as a new answer, thereby concealing the input information such as personal information.

[0043] Examples of image generation AI include, but are not limited to, StackGAN (Generative Adversarial Networks), AttnGAN, T2I (Text-to-Image) with Transformers, DALL-E, etc. Examples of multimodal AI include, but are not limited to, models that generate images from text or generate text from images, such as GPT-4V and CM3Leon (Chameleon Multimodal Model).

[0044] In step S3, the information processing device 1 inputs input information including instruction information, which is information instructing the generation of a first content CA predicted to be accessed by a user U satisfying the classification condition CC, to the generation AI, and causes the generation AI to generate the first content CA. The first content CA is, for example, content shown by at least one of text and image.

[0045] In the example shown in Figure 1, the information processing device 1 pre-stores standard instruction information, and inputs information including the standard instruction information and classification condition information by a worker O to the generation AI as instruction information, thereby causing the generation AI to generate a first content CA for each classification condition CC.

[0046] In the example shown in FIG. 1, the generation AI is a text generation AI, and the standard instruction information is the information of the character string "I will post a certain value. Please create 10 questions for a Q&A site that are likely to be viewed by Japanese users who are looking for or already possess that value. Please do not include too many of the posted values ​​in the sentences. Please output without bullet points or line breaks, as in the following sentences: eg) I want to be socially accepted. What kind of world do you want? Why is the weather bad today?". Also, in the example shown in FIG. 1, the classification condition information is the information of the character string "freedom", and the information processing device 1 adds the information of the character string "values:". As a result, each of the 10 questions that a user U who has the value of freedom is likely to access (e.g., view) is generated as the first content.

[0047] When the generating AI is a GPT, the fixed instruction information is included in the input information (prompt) as an instruction, and the classification condition information is included in the input information (prompt) as a user message, but is not limited to such an example. For example, the input information may include information including the information of the character string "Freedom as a value" as a user message instead of the information of the character string "Post a value. The value" in the fixed instruction information.

[0048] Furthermore, instead of generating a sentence according to the classification condition CC, the information processing device 1 can also generate a keyword group according to the classification condition CC. In this case, the information processing device 1 can use, as the fixed instruction information, information including, for example, a character string "I will post a certain value. Please create 100 keywords contained in questions on a Q&A site that are likely to be viewed by Japanese users who are looking for or already possess that value" and information showing examples of keywords to be output.

[0049] Next, the information processing device 1 classifies each of the multiple second contents CB1 to CBm (m is an integer equal to or greater than 2) into the corresponding classification conditions CC based on the first contents CA for each classification condition CC generated in step S3 (step S4). In the following, when each of the second contents CB1 to CBm is not individually distinguished, it may be referred to as the second content CB. In step S4, the information processing device 1 vectorizes each of the first contents CA and each of the second contents CB.

[0050] The vectorization of the content is performed, for example, by embedding using a language model (for example, a transformer-based model). The vectorized content is represented, for example, by a vector with several hundred dimensions, but is not limited to such an example.

[0051] Embedding using a language model is, for example, embedding using text-embedding-ada provided by OpenAI (registered trademark), BERT (Bidirectional Encoder Representations from Transformers), etc., but is not limited to these examples.

[0052] Note that content vectorization is not limited to embedding using a language model, and content vectorization may be performed using, for example, Doc2Vec, an average of word embedding, etc. Word2Vec, fastText, etc., are used for word embedding.

[0053] The information processing device 1 classifies the second contents CB for each classification condition CC by, for example, extracting, for each classification condition CC, a second content CB whose vector similarity with one or more first contents CA is equal to or greater than a threshold value.

[0054] For example, the information processing device 1 can classify the second content CB into only one classification condition CC, or can classify the second content CB into two or more classification conditions CC. For example, the information processing device 1 can classify the second content CB into only one classification condition CC by classifying the second content CB into the classification condition CC corresponding to the first content CA whose vector similarity is equal to or greater than a threshold and has the highest similarity.

[0055] In addition, when there are two or more first contents CA whose vector similarity with the second content CB is equal to or greater than a threshold value, the information processing device 1 can classify the second content CB into two or more classification conditions CC corresponding to these two or more first contents CA, respectively.

[0056] The similarity of the vectors is a cosine similarity, but may be a Jaccard similarity, or may be the inverse of the Euclidean distance or the inverse of the Manhattan distance. In addition, when the Euclidean distance or the Manhattan distance is used, the information processing device 1 classifies, for example, the second content CB whose Euclidean distance or the Manhattan distance with the first content CA is less than a threshold value, under the classification condition CC corresponding to the first content CA.

[0057] In addition, the information processing device 1 can learn a classification model using vectors and the corresponding hand-labeled data and use the model to perform classification, or can use any method to determine the similarity between vectors or classify vectors, and ultimately the similarity between contents or classify contents.

[0058] In addition, when the first content CA is a group of keywords according to the classification conditions CC, the information processing device 1 calculates a score for each classification condition CC of each second content CB based on the multiple keywords that constitute each first content CA and the multiple characteristic words (e.g., nouns, verbs, adjectives, etc.) contained in each second content CB.

[0059] Then, the information processing device 1 classifies the second content CB into the classification condition CC whose score is equal to or greater than the threshold value. The information processing device 1 can also classify the second content CB into the highest classification condition whose score is equal to or greater than the threshold value.

[0060] For example, the information processing device 1 calculates the number of keywords that match the first content CA according to the classification conditions CC as a score of the classification conditions CC for each classification condition, and classifies the second content CB into the classification conditions CC whose score is equal to or greater than a threshold value.

[0061] The information processing device 1 can also calculate the score of the classification criteria CC by weighting each keyword constituting the first content CA and adding them up. The information processing device 1 can output the weighting for each keyword to, for example, the generation AI.

[0062] For example, in step S3, the information processing device 1 uses information including the string "I will post a certain value. Please create 100 keywords that are contained in questions on a Q&A site that are likely to be viewed by Japanese users who are looking for or already possess that value, along with their importance." as standard instruction information.

[0063] In addition, when the first content CA and the second content CB are images, the information processing device 1 can input information including the first content CA, the second content CB, and instruction information indicating an instruction to output the similarity between the first content CA and the second content CB as input information to the multimodal AI, and output the similarity from the multimodal AI.

[0064] In this case, the information processing device 1 classifies, for example, the content CA into classification conditions CC corresponding to the first content CA having the highest similarity generated by the multimodal AI and a similarity level equal to or greater than a threshold, or into classification conditions CC corresponding to the first content CA having a similarity level equal to or greater than a threshold, generated by the multimodal AI.

[0065] In addition, when the first content CA and the second content CB are images, the information processing device 1 can also cause the multimodal AI to generate an explanatory sentence that explains what images are contained in each of the first content CA and the second content CB.

[0066] In this case, the information processing device 1 vectorizes each of the explanatory text of the first content CA and the explanatory text of the second content CB, and calculates the similarity between the vectorized explanatory text of the first content CA and the vectorized explanatory text of the second content CB. The method of calculating the similarity and the method of classifying the second content CB into the classification condition CC are the same as when the first content CA and the second content CB are text.

[0067] When the first content CA and the second content CB are images, the information processing device 1 can also directly vectorize each of the first content CA and the second content CB to calculate the similarity.

[0068] Furthermore, when the first content CA is a keyword group and the second content CB is an image, the information processing device 1 can also cause the multimodal AI to generate one or more keywords indicating what kind of images are included in the second content CB. In this case, the information processing device 1 uses the one or more keywords generated by the multimodal AI as characteristic words of the second content CB.

[0069] Furthermore, when the first content CA is a keyword group and the second content CB is an image, the information processing device 1 can cause the multimodal AI to generate the relevance between the first content CA and the second content CB. In this case, the information processing device 1 classifies each second content CB into the classification condition CC corresponding to the first content CA whose relevance is equal to or greater than a threshold value.

[0070] In addition, when the first content CA and the second content CB each contain text and an image, the information processing device 1 classifies the first content CA into classification condition CC corresponding to the first content CA whose integrated score, which is a weighted sum of the similarity between the texts and the similarity between the images, is equal to or greater than a threshold, or into classification condition CC corresponding to the first content CA whose integrated score is equal to or greater than the threshold and is the largest.

[0071] Next, the information processing device 1 classifies the users U based on the access information of the users U to the second content CB classified in step S4 (step S5). The access information of the users U to the second content CB includes information such as a user ID (identifier) ​​of the users U who accessed the second content CB, the access date and time of the users U to the second content CB, and the browsing (display) time of the second content CB by the users U. In the example shown in Fig. 1, the multiple users U who accessed the second content CB are indicated as users U1 to Uk (k is an integer equal to or greater than 2).

[0072] The information processing device 1 determines the number of accesses to the second content CB for each classification condition CC based on the access information of the user U. For example, when the time for which the second content CB is browsed (displayed) by the user U is equal to or longer than a threshold, the information processing device 1 determines that the second content CB has been accessed, and counts the number of accesses for each classification condition CC to determine the number of accesses to the second content CB for each classification condition CC.

[0073] Then, the information processing device 1 performs a process of classifying, for each user U, the user U into the classification condition CC that has the largest number of accesses to the second content CB among the number of accesses to the second content CB for each classification condition CC.

[0074] Furthermore, the information processing device 1 can also perform a process of classifying, for each user U, the user U into a classification condition CC in which the number of accesses to the second content CB is equal to or greater than a threshold value, among the number of accesses to the second content CB for each classification condition CC.

[0075] The information processing device 1 can also assign a larger weight to the longer time the second content CB is viewed (displayed), and calculate the sum of the weights for each classification condition CC as the score for each classification condition CC. In this case, the information processing device 1 classifies the user U into a classification condition CC whose score is equal to or greater than a threshold, or into a classification condition CC whose score is the largest.

[0076] Furthermore, the information processing device 1 can also calculate, as the score of the classification condition CC, a value obtained by multiplying the time spent browsing (displaying) the second content CB by a predetermined weight and adding the number of times the user U accessed the second content CB for each classification condition CC. In this case, the information processing device 1 classifies the user U into a classification condition CC whose score is equal to or greater than a threshold, or into a classification condition CC whose score is the largest.

[0077] The information processing device 1 generates a user model based on information of each user U classified for each classification condition CC. The user model is a classification model that receives information of the user U as input and outputs a score for each classification condition CC. Such a classification model is, for example, a Gradient Boosting Decision Tree (GBDT), a random forest, a regression model, a neural network, or the like, but is not limited to these examples.

[0078] The information processing device 1 uses learning information including, for each user U classified under the classification condition CC, information on the user U classified under the classification condition CC and information indicating the classification condition CC under which the user U is classified as described above. The information indicating the classification condition CC under which the user U is classified is used as correct answer data (label). The information on the user U is, for example, attribute information on the user U and the behavior history of the user U, and the values ​​of each item included in this information are used as feature quantities.

[0079] The information processing device 1 inputs information of a user U who is not classified into a classification condition CC as input information to the classification model, and can classify the user U who is not classified into a classification condition CC into a classification condition CC based on the score of each classification condition CC output from the classification model.

[0080] For example, the information processing device 1 can assign the user U to a classification condition CC for which the score output from the classification model is greater than or equal to a threshold, or assign the user U to a classification condition CC for which the score output from the classification model is greater than or equal to a threshold and is the largest.

[0081] In this way, the information processing device 1 acquires classification condition information indicating each of a plurality of classification conditions CC for classifying the user U, and based on the acquired classification condition information, generates, for each classification condition CC, a first content CA used to classify the second content, which is a first content according to the classification condition CC and is accessed by the user U. This allows the information processing device 1 to assist in the generation of the first content CA for classifying the second content CB.

[0082] The configuration of an information processing system including an information processing device 1 that performs such processing, a plurality of terminal devices 2, a terminal device 3, and an information processing device 4 will be described in detail below.

[0083] 2. Configuration of the Information Processing System Fig. 2 is a diagram showing an example of the configuration of an information processing system according to an embodiment. As shown in Fig. 2, the information processing system 100 according to the embodiment includes an information processing device 1, a plurality of terminal devices 2, a terminal device 3, and an information processing device 4.

[0084] The multiple terminal devices 2 are used by different users U, and the terminal device 3 is used by an operator O. The terminal devices 2 and 3 are, for example, a notebook PC (Personal Computer), a desktop PC, a smartphone, a tablet PC, or a wearable device. The wearable device is, for example, a smart glass or a smart watch, but is not limited to such examples.

[0085] The information processing device 4 provides various information to the user U via online sites. Examples of online sites provided by the information processing device 4 include, but are not limited to, a Q&A site, a news site, an EC site, a map site, an image posting site, and a video viewing site.

[0086] The information processing device 1, the terminal device 2, the terminal device 3, and the information processing device 4 are connected to each other via a network N in a wired or wireless manner so as to be able to communicate with each other. Note that the information processing system 100 shown in FIG. 2 may include a plurality of information processing devices 1 and information processing devices 4.

[0087] The network N includes, for example, a Wide Area Network (WAN) such as the Internet and mobile communication networks such as Long Term Evolution (LTE), 4th Generation (4G), and 5th Generation (5G: fifth generation mobile communication system).

[0088] The terminal devices 2, 3 and the information processing device 4 can connect to the network N and communicate with the information processing device 1 via short-range wireless communication such as a mobile communication network, Bluetooth (registered trademark), or a wireless LAN (Local Area Network).

[0089] 3. Configuration of information processing device 1 3 is a diagram showing an example of the configuration of the information processing device 1 according to the embodiment. As shown in FIG. 3, the information processing device 1 includes a communication unit 10, a storage unit 11, and a processing unit 12.

[0090] [3.1. Communication unit 10] The communication unit 10 is realized by, for example, a communication module or a NIC (Network Interface Card). The communication unit 10 is connected to a network N by wire or wirelessly, and transmits and receives information to and from various other devices. For example, the communication unit 10 transmits and receives information to and from each of the terminal device 2, the terminal device 3, and the information processing device 4 via the network N.

[0091] [3.2. Storage section 11] The storage unit 11 is realized by, for example, a semiconductor memory element such as a random access memory (RAM) or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 11 has a user information storage unit 20, a classification information storage unit 21, a first content storage unit 22, a second content storage unit 23, and a classification result storage unit 24.

[0092] 3.2.1. User information storage unit 20 The user information storage unit 20 stores user information including information about the user U. Fig. 4 is a diagram showing an example of a user information table stored in the user information storage unit 20 of the information processing device 1 according to the embodiment. As shown in Fig. 4, the user information table stored in the user information storage unit 20 includes items such as "user ID", "attribute information", and "behavior history".

[0093] The "user ID" is identification information for identifying the user U. The "attribute information" is attribute information of the user U corresponding to the "user ID", and includes, for example, psychographic attribute information and demographic attribute information. The demographic attributes are, for example, gender, age, place of residence, and occupation, and the psychographic attributes are interests such as travel, clothes, cars, and religion, lifestyle, thoughts, and thought trends.

[0094] The "behavior history" includes information on the behavior history of the user U associated with the "user ID." The behavior history of the user U includes, for example, information on the movement history of the user U and information on the behavior history of the user U in online services. The information on the movement history of the user U includes, for example, information on the route traveled by the user U and information on the places visited by the user U.

[0095] The information on the user U's behavior history in the online service includes, for example, the search history and access history of the user U in the online service. The search history of the user U includes, for example, the search history of content by the user U in a web search service. The access history of the user U is, for example, the access history of content by the user U in the online service, and includes information on the content of various sites accessed by the user U.

[0096] [3.2.2. Classification information storage unit 21] The classification information storage unit 21 stores classification information including various information related to classification. Fig. 5 is a diagram showing an example of a classification information table stored in the classification information storage unit 21 of the information processing device 1 according to the embodiment.

[0097] As shown in FIG. 5, the classification information table stored in the classification information storage unit 21 includes items such as "classification ID," "classification condition ID," and "classification condition information." The "classification ID" is identification information of the classification method, and is assigned to each classification method. The classification method identified by the classification ID is, for example, a predetermined classification method or a classification method set by the worker O, but is not limited to such examples. The classification method set by the worker O is a classification method based on multiple classification conditions indicated by multiple classification condition information included in the classification request.

[0098] The "classification condition ID" is identification information of the classification condition used in the classification method corresponding to the "classification ID" and is assigned to each classification condition. The "classification condition information" is information on the classification condition corresponding to the combination of the "classification ID" and the "classification condition ID".

[0099] When the classification conditions indicated in the classification condition information are, for example, classification conditions based on the psychographic attributes of user U, the classification conditions may be, for example, classification conditions based on user U's values, classification conditions based on user U's preferences, classification conditions based on user U's lifestyle, etc., but are not limited to such examples.

[0100] The values ​​of user U are, for example, values ​​classified as freedom, equality, philanthropy, peace, etc., but are not limited to such examples and may be, for example, values ​​classified as tradition, success, self-realization, coexistence, etc., or other values.

[0101] For example, when the object is a car, user U's values ​​may be classified as, for example, emphasis on safety, emphasis on livability, emphasis on design, emphasis on driving performance, etc., and when the object is a house, the values ​​may be classified as, for example, emphasis on location, emphasis on price, emphasis on layout, emphasis on design, emphasis on future potential, etc.

[0102] In addition, the values ​​of user U may be values ​​based on Schwartz's value theory. In this case, the values ​​of user U are classified as power, achievement, hedonism, stimulation, self-determination, universalism, philanthropy, tradition, harmony, and security, and the number of classification conditions is 10 (n=10).

[0103] The lifestyle of the user U may be, for example, classified by the AIO approach, VALS, etc. In the case of classification by the AIO approach, the classification conditions are indicated by one or more items of activities, interests, opinions, etc.

[0104] In the case of classification using VALS, multiple classification conditions are used, for example, self-actualizers, successful people, group members, those who maintain their livelihood, those in poverty, those who want to succeed, those with social conscience, intellectuals, and young intellectuals, and the number of classification conditions is nine (n=9).

[0105] The classification conditions may be, for example, classification conditions shown in the innovator theory that classifies the process of diffusion of new products. In this case, the classification conditions are innovators, early adopters, early majority, late majority, and laggards, and the number of classification conditions is five (n=5).

[0106] 3.2.3. First content storage unit 22 The first content storage unit 22 stores a plurality of first contents generated by the processing unit 12. Fig. 6 is a diagram showing an example of a second content table stored in the first content storage unit 22 of the information processing device 1 according to the embodiment.

[0107] 6, the content table stored in the first content storage unit 22 includes items such as "first content ID," "classification ID," "classification condition ID," and "first content." The "content ID" is identification information of the first content.

[0108] "Classification ID" is the classification ID of the classification method corresponding to "First Content ID." "Classification Condition ID" is the classification condition ID of the classification conditions used in the classification method corresponding to "First Content ID." "First Content" is the first content corresponding to "Content ID."

[0109] 3.2.4. Second content storage unit 23 The second content storage unit 23 stores a plurality of second contents collected by the processing unit 12. Fig. 7 is a diagram showing an example of a second content table stored in the second content storage unit 23 of the information processing device 1 according to the embodiment.

[0110] 7, the content table stored in the second content storage unit 23 includes items such as "second content ID," "site ID," "type ID," "second content," and "access information." The "second content ID" is identification information of the second content.

[0111] The "site ID" is identification information of an online site where the second content corresponding to the "second content ID" is posted or provided, and is assigned to each online site. Examples of the online site include, but are not limited to, a Q&A site, a news site, an EC site, a map site, an image posting site, and a video viewing site.

[0112] The "type ID" is identification information indicating the type of the second content corresponding to the "second content ID." The type of the second content is, for example, a news article on a news site, a comment on a news article on a news site, a product description on a product page on an EC site, a question on a Q&A site, an image content on an image posting site, a title and a caption text on a video viewing site, and the like, but is not limited to these examples.

[0113] The "second content" is the second content corresponding to the "second content ID." The "access information" is access information of the user U to the second content corresponding to the "second content ID." Such access information includes, for example, information such as the user ID of the user U who accessed the second content, the date and time of the user U's access to the second content, and the viewing (display) time of the user U of the second content.

[0114] [3.2.5. Classification result storage unit 24] The classification result storage unit 24 stores the classification result of the second content by the processing unit 12. Fig. 8 is a diagram showing an example of a classification result table stored in the classification result storage unit 24 of the information processing device 1 according to the embodiment.

[0115] 8, the classification result table stored in the classification result storage unit 24 includes items such as "second content ID," "classification result," etc. The "second content ID" is identification information of the second content.

[0116] The "classification result" includes information indicating the classification result of the second content corresponding to the "second content ID" by the processing unit 12. The information indicating the classification result includes, for example, a combination of a classification ID and a classification condition ID for each classification by the processing unit 12.

[0117] [3.3. Processing section 12] The processing unit 12 is a controller, and is realized by a processor such as a CPU (Central Processing Unit) or MPU (Micro Processing Unit) by executing various programs (corresponding to an example of an information processing program) stored in a storage device inside the information processing device 1 using a RAM or the like as a working area.

[0118] Furthermore, the processing unit 12 is a controller, and may be realized by an integrated circuit such as an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or a general purpose graphic processing unit (GPGPU).

[0119] 3, the processing unit 12 has an acquisition unit 30, a collection unit 31, a reception unit 32, an extraction unit 33, a determination unit 34, a generation unit 35, a first classification unit 36, a second classification unit 37, a third classification unit 38, and a provision unit 39, and realizes or executes the functions and actions of information processing described below. Note that the internal configuration of the processing unit 12 is not limited to the configuration shown in FIG. 3, and may be other configurations as long as they perform the information processing described below.

[0120] [3.3.1. Acquisition part 30] The acquisition unit 30 acquires various pieces of information. For example, the acquisition unit 30 acquires various pieces of information from the terminal device 2, the terminal device 3, etc. via the communication unit 10.

[0121] For example, the acquisition unit 30 acquires user information transmitted from an external information processing device or the terminal device 2, and updates the user information table stored in the user information storage unit 20 based on the acquired user information.

[0122] Furthermore, the acquisition unit 30 acquires information included in various requests accepted by the acceptance unit 32. For example, the acquisition unit 30 acquires, as classification condition information, information indicating each of a plurality of classification conditions included in the classification request accepted by the acceptance unit 32. Furthermore, the acquisition unit 30 acquires information indicating the type of target content included in the classification request accepted by the acceptance unit 32.

[0123] Furthermore, when the classification request accepted by the accepting unit 32 includes information indicating the classification method or information indicating example sentences instead of information indicating the classification conditions, the acquiring unit 30 acquires the information indicating the classification method or the information indicating example sentences included in the classification request. Then, based on the information indicating the classification method included in the classification request, the acquiring unit 30 acquires, as classification condition information, information indicating each of the multiple classification conditions used for the classification method specified in the classification request from the storage unit 11 or the like.

[0124] Furthermore, the acquisition unit 30 acquires various information from the storage unit 11. The acquisition unit 30 acquires user information, classification information, first content, second content, classification results, classification models, and the like from the storage unit 11. For example, the acquisition unit 30 acquires from the storage unit 11 classification condition information indicating each of a plurality of classification conditions for classifying the user U.

[0125] [3.3.2. Collection unit 31] The collection unit 31 collects various information from the information processing device 4 and the like. For example, the collection unit 31 collects a plurality of second contents, stores the collected plurality of second contents in the second content storage unit 23, and updates the second content table.

[0126] The collection unit 31 collects, for example, a plurality of second contents corresponding to a target content type identified in the classification request received by the reception unit 32. For example, when the target content type identified in the classification request received by the reception unit 32 is a question on a Q&A site, the collection unit 31 collects the question on the Q&A site as the second content.

[0127] In addition, the collection unit 31 collects a plurality of second contents and access information for each of the plurality of second contents, and stores the collected plurality of second contents and their access information in the second content storage unit 23 to update the second content table.

[0128] The second content is, for example, posted content posted to an online site provided by the information processing device 4. The online site is, for example, a Q&A site, an EC site, a news site, an image posting site, a video viewing site, or the like.

[0129] If the online site is a Q&A site, the posted content is a question, but is not limited to this example and may be, for example, an answer. If the online site is an EC site, the posted content is a product description, but is not limited to this example and may be, for example, a product review.

[0130] Furthermore, if the online site is a news site, the posted content may be a news article or a comment on a news article, but is not limited to such an example.

[0131] In addition, if the online site is an image posting site, the posted content is, but is not limited to, image content. In addition, if the online site is an image posting site, the posted content is, but is not limited to, title and caption text, for example.

[0132] [3.3.3. Reception unit 32] The reception unit 32 receives various requests and various pieces of information from the terminal device 2, the terminal device 3, an external information processing device, or the like via the communication unit .

[0133] For example, the reception unit 32 receives information indicating the classification conditions. For example, the reception unit 32 receives a classification request transmitted from the terminal device 3, thereby receiving information indicating each classification condition indicated in the classification request, information indicating the type of target content, and the like.

[0134] The classification request includes information indicating a plurality of classification conditions. For example, information on a character string "freedom", information on a character string "equality", information on a character string "philanthropy", and information on a character string "peace" are each included in the classification request as information indicating a classification condition.

[0135] The classification request may also include, for example, information on the character string "tradition", information on the character string "success", information on the character string "self-realization", and information on the character string "symbiosis" as information indicating the classification conditions.

[0136] In addition, the classification request may include, for example, information indicating the classification conditions, such as the character string "emphasis on safety," the character string "emphasis on livability," the character string "emphasis on design," and the character string "emphasis on driving performance."

[0137] In addition, the classification request may include, for example, information indicating the classification conditions, such as the character string "emphasis on location," the character string "emphasis on price," the character string "emphasis on layout," the character string "emphasis on design," and the character string "emphasis on future potential."

[0138] The classification request also includes information indicating the type of target content selected or specified by the worker O. The type of target content is, for example, a type of second content, and is indicated, for example, by a combination of an online site and a content type. For example, the type of target content is, but is not limited to, a news article on a news site, a comment on a news article on a news site, a product description on a product page on an EC site, a question on a Q&A site, image content on an image posting site, and a title and caption text of video content on a video viewing site.

[0139] Also, instead of the information indicating the classification conditions, a classification request including information indicating each of the classification method, the target content type, and the example sentence selected by the operator O may be accepted. Note that the example sentence does not have to be included in the classification request.

[0140] The classification method selected by the worker O may be, for example, a classification method based on multiple classification conditions included in past classification requests, a classification method based on values ​​based on Schwartz's value theory, a classification method based on the AIO approach, a classification method based on VALS, a classification method based on classification conditions shown in the innovator theory, etc., but is not limited to such examples. Note that example sentences may not be included.

[0141] [3.3.4. Extraction part 33] The extraction unit 33 extracts, from the plurality of second contents collected by the collection unit 31, second contents whose access information satisfies a predetermined condition.

[0142] For example, the extraction unit 33 extracts, from among the plurality of second contents collected by the collection unit 31, second contents whose access information satisfies a predetermined condition for each type of second content.

[0143] The type of the second content may be, for example, if the online site is a Q&A site, the second content is a question; if the online site is an e-commerce site, the second content is a product description; if the online site is a news site, the second content is a news article, image content on an image posting site, or the title or caption text of video content on a video viewing site, but is not limited to such examples.

[0144] The predetermined condition is, for example, a condition that the number of accesses per unit time period or the total number of accesses is equal to or greater than a threshold, but is not limited to such an example. For example, the predetermined condition may be a condition that the number of accesses per unit time period or the total number of accesses is equal to or greater than a threshold and the degree of evaluation is equal to or greater than a threshold, or may be another condition.

[0145] The predetermined condition may be a condition that the distribution of attributes of the accessed users U is highly biased. For example, the highly biased distribution may be a case where the access ratio of users U having a specific combination of attribute item values ​​(e.g., men in their 20s or women in their 30s) is equal to or greater than a threshold value (e.g., 30%), but is not limited to such an example.

[0146] [3.3.5. Decision Unit 34] The determining unit 34 determines a plurality of classification conditions based on the plurality of second contents collected by the collecting unit 31.

[0147] For example, the determination unit 34 uses the generation AI to determine the values ​​that a user viewing each second content has or is likely to have, and determines a plurality of classification conditions based on the determination result. The determination unit 34 uses information including the standard instruction information and the second content as input information for the generation AI, and causes the generation AI to output information indicating the values, thereby obtaining the information.

[0148] The standard instruction information is, for example, a character string "Post content. Please output the values ​​that users who view the content have or are likely to have in the following format: eg) Values: Freedom".

[0149] The determination unit 34 determines, as a classification condition, each of two or more values ​​whose number determined by the generation AI is equal to or greater than a threshold value, based on the values ​​of the users viewing each second content acquired from the generation AI. The determination unit 34 can also determine, as a classification condition, each of the top p values ​​(p is an integer equal to or greater than 2) whose number determined by the generation AI is the largest.

[0150] The determination unit 34 also inputs input information to the generation AI, including instruction information indicating an instruction to organize and aggregate the user's values ​​into 10 or less values ​​and information indicating the values ​​of the user viewing each second content acquired from the generation AI, and determines each of the 10 or less values ​​output from the generation AI as a classification condition. In this case, the instruction information includes, for example, the character string "Please organize the following multiple values, aggregate them into 10 or less values, and output them."

[0151] In addition, the determination unit 34 can also cause the generation AI to determine a plurality of classification conditions based on the results of summarizing each second content or converting it into a plurality of keywords. The determination unit 34 can cause the generation AI to summarize the second content or convert it into a plurality of keywords.

[0152] For example, the determination unit 34 uses information including the string "Please summarize the content below in 50 characters or less" or "Please extract keywords from the content below in terms of values ​​or less" and the second content as input information for the generation AI, and causes the generation AI to summarize the second content or convert it into multiple keywords.

[0153] Then, the determination unit 34 obtains information indicating the values ​​by having the generation AI output information including, for example, a string of information "Please output 10 values ​​that are imagined from the following group of summaries or group of keywords." and information on summaries and group of keywords of each second content as input information to the generation AI.

[0154] It should be noted that the determining unit 34 can determine classification conditions based on lifestyles or other classification conditions instead of or in addition to values, based on the plurality of second contents collected by the collecting unit 31.

[0155] [3.3.6. Generation unit 35] The generating unit 35 generates various contents and models. For example, the generating unit 35 generates first contents according to each classification condition based on the classification condition information for each classification condition acquired by the acquiring unit 30. The first contents are contents used to classify second contents accessed by the user U.

[0156] For example, the generating unit 35 generates a first content according to each classification condition, based on the classification condition information for each classification condition acquired by the acquiring unit 30 and information indicating the type of target content.

[0157] The classification conditions indicated in the classification condition information are, for example, classification conditions based on the psychographic attributes of user U, such as, as described above, classification conditions based on user U's values, classification conditions based on user U's preferences, classification conditions based on user U's lifestyle, etc., but are not limited to such examples.

[0158] For example, the lifestyle of user U is classified based on the behavioral patterns of user U, such as classifications such as outdoorsy or indoor, classifications such as morning type, night type, or something in between, classifications such as desk work, fieldwork, or something in between, but is not limited to such examples.

[0159] The generating unit 35 generates, for each classification condition, a content predicted to be accessed by a user U who satisfies the classification condition indicated by the classification condition information as a first content. For example, when the classification condition indicated by the classification condition information is the values ​​of the user U, the generating unit 35 generates, for each classification condition, a content predicted to be accessed by a user U who has the values ​​indicated by the classification condition as a first content.

[0160] As mentioned above, the target content types include, for example, news articles on a news site, comments on news articles on a news site, product descriptions on product pages on e-commerce sites, questions on Q&A sites, image content on image posting sites, and titles and caption text of video content on video viewing sites, but are not limited to such examples.

[0161] In addition, when the classification request includes information indicating the classification method selected by the worker O, the generation unit 35 generates content that is predicted to be accessed by the user U who satisfies the classification condition CC corresponding to the classification method selected by the worker O as a first content CA for each classification condition CC.

[0162] For example, the generating unit 35 generates the first content according to each classification condition based on the classification conditions used in the classification method specified in the classification request and the information indicating the example sentences. Note that, when the second content to be classified is a specific type of content of a specific site, the information indicating the target content type may not be used.

[0163] The generation unit 35 generates the first content for each classification condition by using, for example, a generation AI. The generation AI is, for example, a text generation AI, an image generation AI, a multimodal AI, etc. The generation AI is, for example, disposed in an external information processing device, but is not limited to such an example. For example, the generation AI may be disposed in the generation unit 35.

[0164] The generation unit 35 inputs input information including information instructing the generation AI to generate a first content, which is a content predicted to be accessed by a user who satisfies the classification condition, and causes the generation AI to generate the first content. The generation unit 35 inputs the input information to the generation AI via an API provided by an external information processing device, for example, and causes the generation AI to generate and acquire the first content for each classification condition.

[0165] A text generation AI is, for example, a language model trained to predict and output the next token from an input token sequence, such as a transfer-based model or an RNN-based model.

[0166] Examples of transfer-based models include, but are not limited to, GPT and BARD. Examples of RNN-based models include, but are not limited to, RWKV. It is preferable that the input information is trained so that it is not used as a new answer, thereby concealing the input information such as personal information.

[0167] Examples of image generation AI include, but are not limited to, StackGAN, AttnGAN, T2I with Transformers, DALL-E, etc. Examples of multimodal AI include, but are not limited to, models that generate images from text or generate text from images, such as CM3Leon, etc.

[0168] The generation unit 35 inputs input information including instruction information, which is information instructing the generation of content predicted to be accessed by the user U who satisfies the classification condition, to the generation AI, and causes the generation AI to generate the first content. The first content is, for example, content shown by at least one of text and image.

[0169] The second content is, for example, posted content posted on an online site, and the generating unit 35 generates, as the first content, posted content that is predicted to be accessed by a user U who has the values ​​indicated by the classification conditions.

[0170] For example, if the online site is a Q&A site, the second content is a question; if the online site is an e-commerce site, the second content is a product description; if the online site is a news site, the second content is a news article, but is not limited to such examples.

[0171] The generation unit 35, for example, stores standard instruction information in advance, and inputs information including the standard instruction information and classification condition information by worker O to the generation AI as instruction information, thereby causing the generation AI to generate the first content for each classification condition.

[0172] Here, an example will be described in which the classification method is set by the worker O and the target content type is a question on a Q&A site. Fig. 9 is a diagram showing an example of input information input to the generation AI by the generation unit 35 of the information processing device 1 according to the embodiment.

[0173] In the example shown in FIG. 9, the generation AI is a text generation AI, and the standard instruction information is the information of the character string "I will post a certain value. Please create 10 questions for a Q&A site that are likely to be viewed by Japanese users who are looking for or already possess that value. Please do not include too many of the posted values ​​in the sentences. Please output without bullet points or line breaks, as in the following sentences. eg) I want to be socially accepted. What kind of world do you want? Why is the weather bad today?". Also, in the example shown in FIG. 9, the classification condition information is the information of the character string "freedom", and the information of the character string "values:" is added by the generation unit 35. As a result, each of the 10 questions that a user U who has the value of freedom is likely to access (e.g., view) is generated as the first content.

[0174] When the generation AI is a GPT, the standard instruction information is included in the input information (prompt) as an instruction, and the classification condition information is included in the input information (prompt) as a user message, but is not limited to such an example.

[0175] For example, the input information may include information including the character string "Freedom as a value" as a user message instead of the character string "Post a certain value. That value" in the fixed instruction information.

[0176] Furthermore, instead of generating sentences according to the classification conditions, the generating unit 35 can generate a keyword group according to the classification conditions. In this case, the generating unit 35 can use, as the fixed instruction information, information including, for example, a character string "I will post a certain value. Please create 100 keywords contained in questions on a Q&A site that are likely to be viewed by Japanese users who are looking for or already possess that value" and information showing examples of keywords to be output.

[0177] In addition, the generation unit 35 can cause the generation AI to generate the first content by using, as input information, information further including, as reference information, the second content extracted by the extraction unit 33. The second content extracted by the extraction unit 33 is the second content that satisfies the above-mentioned predetermined condition.

[0178] For example, the generation unit 35 causes the generation AI to generate the first content using input information including the standard instruction information, which includes the string "Please refer to the following question as a question for the Q&A site," and the second content extracted by the extraction unit 33 as reference information.

[0179] Next, an example will be described in which the classification method is the classification condition indicated by the innovator theory and the target content type is an EC site. Fig. 10 is a diagram showing another example of input information input to the generation AI by the generation unit 35 of the information processing device 1 according to the embodiment.

[0180] In the example shown in FIG. 10, the generation AI is a text generation AI or a multimodal AI, and the standard instruction information is the information of the character string "Post a certain user type. Please create 10 sentences of product description for an EC site that is likely to be viewed by Japanese users who are looking for or already have that user type. Do not include the posted user type in the sentences as they are. Please output without bullet points or line breaks, as in the following sentence. eg) We use cutting-edge technology. Safety and security are ensured. We have achieved the highest sales in Japan." Also, in the example shown in FIG. 10, the classification condition information is information of the character string "innovator", and the information of the character string "user type:" is added by the generation unit 35.

[0181] In this case, the generation unit 35 can also cause the generation AI to generate the first content by using, as input information, information further including, as reference information, the second content extracted by the extraction unit 33. The second content extracted by the extraction unit 33 is the second content that satisfies the above-mentioned predetermined condition.

[0182] Next, an example will be described in which the classification method is the classification condition indicated by the innovator theory, and the target content type is image content (for example, still image content or video content) of an image posting site. Fig. 11 is a diagram showing yet another example of input information input to the generation AI by the generation unit 35 of the information processing device 1 according to the embodiment.

[0183] In the example shown in Fig. 11, the generation AI is an image generation AI or a multimodal AI, and the standard instruction information is a character string "I will post a certain user type. Please create 10 images for an image posting site that are likely to be viewed by Japanese users who are looking for that user type or who already have that user type." Also, in the example shown in Fig. 11, the classification condition information is a character string "innovator," and the generation unit 35 adds the character string "user type:."

[0184] In this case, the generation unit 35 can also cause the generation AI to generate the first content by using, as input information, information further including, as reference information, the second content extracted by the extraction unit 33. The second content extracted by the extraction unit 33 is the second content that satisfies the above-mentioned predetermined condition.

[0185] Next, an example will be described in which the classification method is the classification condition indicated by the innovator theory, and the target content type is the title and caption text of a video content on a video browsing site. Fig. 12 is a diagram showing another example of input information input to the generation AI by the generation unit 35 of the information processing device 1 according to the embodiment.

[0186] In the example shown in FIG. 12, the generation AI is a text generation AI or a multimodal AI, and the standard instruction information is the character string "Post a certain user type. Please create 10 sentences of title and caption text for video content on a video viewing site that is likely to be viewed by Japanese users who are looking for or already have that user type. Do not include the posted user type in the sentences as they are. The output should be output without bullet points or line breaks, as in the following sentence. eg) Coach De Zerbi's comments on the in-form attacker. Real Madrid turns their backs on Kubo Takefusa's heroic performance." Also, in the example shown in FIG. 12, the classification condition information is the character string "innovator", and the generation unit 35 adds the character string "user type:".

[0187] In this case, the generation unit 35 can also cause the generation AI to generate the first content by using, as input information, information further including, as reference information, the second content extracted by the extraction unit 33. The second content extracted by the extraction unit 33 is the second content that satisfies the above-mentioned predetermined condition.

[0188] Next, an example will be described in which the classification method is a classification condition based on the lifestyle of the user U, and the target content type is a news article on a news site. Fig. 13 is a diagram showing yet another example of input information input to the generation AI by the generation unit 35 of the information processing device 1 according to the embodiment.

[0189] In the example shown in FIG. 13, the generation AI is a text generation AI or a multimodal AI, and the standard instruction information is the character string "Post a certain lifestyle. Please create 10 sentences of news article titles that are likely to be viewed by Japanese users who are looking for or already have that lifestyle. Do not include too much of the lifestyle you posted in the sentences. Please output without bullet points or line breaks, as in the following sentence: eg) A Bon Odori festival will be held on September 19th. The Japanese national soccer team wins. The Prime Minister visits the United States." Also, in the example shown in FIG. 13, the classification condition information is the character string "People who spend time on work," and the generation unit 35 adds the character string "Lifestyle:" information.

[0190] Moreover, the generation unit 35 generates a user model based on the information of each user U classified for each classification condition by the first classification unit 36. The user model is a classification model that receives the information of the user U as input and outputs a score for each classification condition. Such a classification model is, for example, GBDT, random forest, regression model, neural network, etc., but is not limited to these examples.

[0191] The generating unit 35 generates a user model using learning information including, for each user U classified into a classification condition, information on the user U classified into the classification condition and information indicating the classification condition into which the user U was classified as described above. The information indicating the classification condition into which the user U was classified is used as correct answer data (label). The information on the user U is, for example, attribute information on the user U and the behavior history of the user U, and the values ​​of each item included in this information are used as feature quantities.

[0192] [3.3.7. 1st classification section 36] The first classification unit 36 ​​classifies each of the second contents based on the first contents for each classification condition generated by the generation unit 35. When classifying the second contents, the first classification unit 36 ​​performs, for example, vectorization of each of the first contents and vectorization of each of the second contents.

[0193] The content is vectorized, for example, by embedding with a language model (e.g., a transformer-based model). The vectorized content is represented, for example, as a vector with several hundred dimensions, but is not limited to such an example.

[0194] Embedding by a language model is, for example, embedding by text-embedding-ada provided by OpenAI, Inc., BERT, etc., but is not limited to such examples. Note that content vectorization is not limited to embedding by a language model, and content vectorization may be performed by, for example, Doc2Vec, the average of word embedding, etc. For word embedding, for example, Word2Vec, fastText, etc. are used.

[0195] For example, the first classification unit 36 ​​extracts, for each classification condition, second content whose vector similarity with one or more first contents is equal to or greater than a threshold, and classifies each of the multiple second contents according to the corresponding classification condition.

[0196] For example, the first classification unit 36 ​​can classify the second content into only one classification condition, or can classify the second content into two or more classification conditions. For example, the generation unit 35 can classify the second content into only one classification condition by classifying the second content into a classification condition corresponding to the first content having the highest similarity, the similarity of the vector of the second content being equal to or greater than a threshold value.

[0197] In addition, when there are two or more first contents whose vector similarity with the second content is equal to or greater than a threshold, the first classification unit 36 ​​can classify the second content into two or more classification conditions corresponding to these two or more first contents, respectively.

[0198] The similarity of the vectors is a cosine similarity, but may be a Jaccard similarity, or may be the inverse of the Euclidean distance or the inverse of the Manhattan distance. When the Euclidean distance or the Manhattan distance is used, the first classification unit 36 ​​classifies, for example, a second content whose Euclidean distance or the Manhattan distance with the first content is less than a threshold value, according to the classification condition corresponding to the first content.

[0199] In addition, the first classification unit 36 ​​can also use any method to learn a classification model using vectors and the corresponding hand-labeled data and use it to perform classification, or to identify the similarity between vectors or classify vectors, and ultimately the similarity between contents or classify contents.

[0200] In addition, when the first content is a group of keywords according to the classification conditions, the first classification unit 36 ​​calculates a score for each classification condition of each second content based on the multiple keywords that constitute each first content and the multiple characteristic words (e.g., nouns, verbs, adjectives, etc.) included in each second content.

[0201] Then, the first classification unit 36 ​​classifies the second content into the classification condition for which the score is equal to or greater than the threshold. The first classification unit 36 ​​can also classify the second content into the highest classification condition for which the score is equal to or greater than the threshold.

[0202] For example, the first classification unit 36 ​​calculates the number of keywords that match the first content according to the classification condition as a score of the classification condition, and classifies the second content into the classification condition whose score is equal to or greater than a threshold value.

[0203] The first classification unit 36 ​​can also calculate the score of the classification condition by weighting each keyword constituting the first content and adding the weights. The first classification unit 36 ​​can output the weight for each keyword to, for example, the generation AI.

[0204] For example, the first classification unit 36 ​​uses information including the string "I will post a certain value. Please create 100 keywords that are contained in questions on a Q&A site that are likely to be viewed by Japanese users who are looking for or already possess that value, along with their importance." as standard instruction information.

[0205] In addition, when the first content and the second content are images, the first classification unit 36 ​​can input information including the first content, the second content, and instruction information indicating an instruction to output the similarity between the first content and the second content as input information to the multimodal AI, and cause the multimodal AI to output the similarity.

[0206] In this case, the first classification unit 36 ​​classifies, for example, into a classification condition corresponding to the first content having the highest similarity generated by the multimodal AI, which has a similarity level equal to or higher than a threshold, or into a classification condition corresponding to the first content having a similarity level equal to or higher than a threshold, which is generated by the multimodal AI.

[0207] In addition, when the first content and the second content are images, the first classification unit 36 ​​can also cause the multimodal AI to generate an explanatory sentence that explains what images are contained in each of the first content and the second content.

[0208] In this case, the first classification unit 36 ​​vectorizes each of the explanatory text of the first content and the explanatory text of the second content, and calculates the similarity between the vectorized explanatory text of the first content and the vectorized explanatory text of the second content. The method of calculating the similarity and the method of classifying the second content into the classification conditions are the same as when the first content and the second content are text. When the first content and the second content are images, the first classification unit 36 ​​can also directly vectorize each of the first content and the second content to calculate the similarity.

[0209] Furthermore, when the first content is a keyword group and the second content is an image, the first classification unit 36 ​​can also cause the multimodal AI to generate one or more keywords indicating what kind of images are included in the second content. In this case, the first classification unit 36 ​​uses the one or more keywords generated by the multimodal AI as characteristic words of the second content.

[0210] In addition, when the first content is a group of keywords and the second content is an image, the first classification unit 36 ​​can cause the multimodal AI to generate the relevance between the first content and the second content. In this case, the first classification unit 36 ​​classifies each second content according to a classification condition corresponding to the first content whose relevance is equal to or greater than a threshold value.

[0211] In addition, when the first content and the second content each include text and an image, the first classification unit 36 ​​classifies the first content into a classification condition corresponding to the first content whose integrated score, which is a weighted sum of the similarity between the texts and the similarity between the images, is equal to or greater than a threshold, or into a classification condition corresponding to the first content whose integrated score is equal to or greater than the threshold and is the largest.

[0212] [3.3.8.Second classification section 37] The second classification unit 37 classifies the users U based on the access information of the users U to the second content classified by the first classification unit .

[0213] The access information of the user U to the second content includes information such as the user ID of the user U who accessed the second content, the date and time when the user U accessed the second content, and the viewing (display) time of the user U to the second content.

[0214] The second classification unit 37 determines the number of accesses to the second content for each classification condition based on the access information of the user U. For example, if the time for which the second content was viewed (displayed) by the user U is equal to or longer than a threshold, the second classification unit 37 determines that the second content has been accessed, and counts the number of accesses for each classification condition to determine the number of accesses to the second content for each classification condition.

[0215] Then, the second classification unit 37 performs a process of classifying, for each user U, the user U into a classification condition that has the largest number of accesses to the second content among the number of accesses to the second content for each classification condition.

[0216] Furthermore, the second classification unit 37 can also perform a process of classifying, for each user U, the user U into a classification condition in which the number of accesses to the second content is equal to or greater than a threshold value, among the number of accesses to the second content for each classification condition.

[0217] The second classification unit 37 can also assign a larger weight to a content viewed (displayed) time and calculate the sum of the weights for each classification condition as a score for each classification condition. In this case, the second classification unit 37 classifies the user U into a classification condition whose score is equal to or greater than a threshold value, or into a classification condition whose score is the largest.

[0218] The second classification unit 37 can also calculate, as the score of the classification condition, a value obtained by multiplying the time spent viewing (displaying) the content by a predetermined weight and adding the number of times the user U accessed the content for each classification condition. In this case, the second classification unit 37 classifies the user U into a classification condition whose score is equal to or greater than a threshold value, or into a classification condition whose score is the largest.

[0219] [3.3.9. Third classification section 38] The third classification unit 38 classifies a user U who is not classified into the classification conditions into the classification conditions. For example, the third classification unit 38 classifies a user U who has no access history to the second content or a user U who was not classified by the second classification unit 37 into the classification conditions.

[0220] The third classification unit 38, for example, inputs information of a user U who is not classified into a classification condition as input information to the classification model, and can classify the user U who is not classified into a classification condition into a classification condition based on the score of each classification condition output from the classification model.

[0221] For example, the second classification unit 37 can assign user U to a classification condition for which the score output from the classification model is greater than or equal to a threshold, or assign user U to a classification condition for which the score output from the classification model is greater than or equal to a threshold and is the largest.

[0222] [3.3.10.Providing Department 39] The providing unit 39 provides the classification result to the worker O by transmitting the classification result, which is information indicating the classification conditions of each user U classified by the second classification unit 37 and information indicating the classification conditions of each user U classified by the third classification unit 38, to the terminal device 3.

[0223] In addition, the providing unit 39 can provide the worker O with information on the user model for each classification condition by transmitting the information on the user model for each classification condition generated by the generating unit 35 to the terminal device 3.

[0224] Furthermore, the providing unit 39 can provide the terminal device 3 with selection information for selecting a classification method, a target content type, example sentences, etc. The selection information includes information on a plurality of classification methods, information on a plurality of target content types, and information on a plurality of example sentences, and each of the classification methods, target content types, and example sentences is displayed on the terminal device 3 in a selectable manner.

[0225] Fig. 14 is a diagram showing an example of a classification request setting screen displayed on the terminal device 3 based on the selection information transmitted by the providing unit 39 of the information processing device 1 according to the embodiment. A classification request setting screen 50 shown in Fig. 14 includes a classification method selection box 51, a target content type selection box 52, and a classification request button 53.

[0226] The classification method selection box 51 is a UI for selecting a classification method, and is, for example, a combo box, but may also be a check box or a drop-down list. FIG. 13 is a diagram showing another example of input information input to the generation AI by

[0227] The target content type selection box 52 is a UI for selecting a target content type, and is, for example, a combo box, but may also be a check box or a drop-down list.

[0228] The classification request button 53 is a button for transmitting a classification request including information indicating the classification method selected in the classification method selection box 51 and information indicating the target content type selected in the target content type selection box 52 to the information processing device 1.

[0229] By operating the terminal device 3, the worker O selects a classification method and a target content type in the classification method selection box 51 and the target content type selection box 52, and selects the classification request button 53, whereby a classification request including the selection results of the worker O is sent from the terminal device 3 to the information processing device 1.

[0230] 4. Processing Procedure Next, a procedure of information processing by the processing unit 12 of the information processing device 1 according to the embodiment will be described. Fig. 15 is a flowchart showing an example of information processing by the processing unit 12 of the information processing device 1 according to the embodiment.

[0231] 15, the processing unit 12 of the information processing device 1 determines whether or not it is time to classify (step S10). The classification timing is, for example, the timing when the classification request transmitted from the terminal device 3 is accepted by the information processing device 1, or the timing that arrives at a predetermined cycle.

[0232] When the processing unit 12 determines that the classification timing has arrived (step S10: Yes), the processing unit 12 collects a plurality of second contents to be classified from the information processing device 4 (step S11). Then, the processing unit 12 determines whether or not there is a classification condition (step S12).

[0233] For example, in step S12, when a classification request transmitted from the terminal device 3 is accepted by the information processing device 1, the processing unit 12 determines that a classification condition exists if the classification request includes information indicating a classification method or classification condition information. If the classification request includes information indicating a classification method, the classification condition is the classification condition of the classification method, and if the classification request includes classification condition information, the classification condition is the classification condition indicated by the classification condition information.

[0234] When it is determined that there is no classification condition (step S12: No), the processing unit 12 determines a plurality of classification conditions based on a plurality of second contents (step S13). When the processing unit 12 ends the processing of step S13, when it is determined that there is a classification condition (step S12: Yes), or when it is determined that the classification timing has not come (step S10: No), the processing unit 12 performs classification processing (step S14). The classification processing of step S14 is the processing of steps S20 to S22 shown in FIG. 16, and will be described in detail later.

[0235] When the process of step S14 is completed, the processing unit 12 determines whether or not the operation end timing has arrived (step S15). The processing unit 12 determines that the operation end timing has arrived when, for example, the power supply of the information processing device 1 is turned off.

[0236] If the processing unit 12 determines that the operation end time has not arrived (step S15: No), it transitions to step S10, and if it determines that the operation end time has arrived (step S15: Yes), it terminates the processing shown in Figure 15.

[0237] Fig. 16 is a flowchart showing an example of classification processing by the processing unit 12 of the information processing device 1 according to the embodiment. As shown in Fig. 16, the processing unit 12 generates first content for each classification condition based on a plurality of classification conditions (step S20).

[0238] Next, the processing unit 12 classifies the second contents for each classification condition based on the first contents for each classification condition generated in step S20 (step S21). Then, the processing unit 12 classifies the user U into the classification condition based on the access information to the second contents of the user (step S22), and ends the process shown in FIG.

[0239] 5. Modifications The generation unit 35 can also cause the generation AI to generate vectorized first content. In this case, the first classification unit 36 ​​does not vectorize the first content, but classifies the second content using the vectorized first content generated by the generation unit 35.

[0240] [6. Hardware Configuration] The information processing device 1 according to the embodiment described above is realized by a computer 80 having a configuration as shown in Fig. 17. Fig. 17 is a hardware configuration diagram showing an example of the computer 80 that realizes the functions of the information processing device 1 according to the embodiment. The computer 80 has a CPU 81, a RAM 82, a ROM (Read Only Memory) 83, a HDD (Hard Disk Drive) 84, a communication interface (I / F) 85, an input / output interface (I / F) 86, and a media interface (I / F) 87.

[0241] The CPU 81 operates and controls each part based on a program stored in the ROM 83 or the HDD 84. The ROM 83 stores a boot program executed by the CPU 81 when the computer 80 is started up, and programs that depend on the hardware of the computer 80.

[0242] The HDD 84 stores programs executed by the CPU 81, data used by such programs, etc. The communication interface 85 receives data from other devices via the network N (see FIG. 2) and sends it to the CPU 81, and transmits data generated by the CPU 81 to other devices via the network N.

[0243] The CPU 81 controls output devices such as a display and a printer, and input devices such as a keyboard and a mouse, via the input / output interface 86. The CPU 81 acquires data from the input devices via the input / output interface 86. The CPU 81 also outputs data generated via the input / output interface 86 to the output devices.

[0244] The media interface 87 reads a program or data stored in a recording medium 88 and provides it to the CPU 81 via the RAM 82. The CPU 81 loads the program from the recording medium 88 onto the RAM 82 via the media interface 87 and executes the loaded program. The recording medium 88 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or a PD (Phase change rewritable Disc), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory.

[0245] For example, when the computer 80 functions as the information processing device 1 according to the embodiment, the CPU 81 of the computer 80 executes a program loaded onto the RAM 82 to realize the function of the processing unit 12. In addition, the HDD 84 stores data in the storage unit 11. The CPU 81 of the computer 80 reads and executes these programs from a recording medium 88, but as another example, these programs may be obtained from another device via the network N.

[0246] [7. Other] In addition, among 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 by 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 arbitrarily unless otherwise specified. For example, the various information shown in each drawing is not limited to the illustrated information.

[0247] In addition, each component of each device shown in the figure is a functional concept, and does not necessarily have to be physically configured as shown in the figure. In other words, the specific form of distribution and integration of each device is not limited to that shown in the figure, and all or part of them can be functionally or physically distributed and integrated in any unit according to various loads, usage conditions, etc.

[0248] For example, the above-mentioned information processing device 1 may be realized by a terminal device and a server computer, or may be realized by multiple server computers. Also, depending on the functions, the configuration can be flexibly changed, such as by calling an external platform using an API or network computing.

[0249] Furthermore, the above-described embodiments and modifications can be appropriately combined as long as the processing contents are not contradictory.

[0250] 8. Effects As described above, the information processing device 1 according to the embodiment includes the acquisition unit 30 and the generation unit 35. The acquisition unit 30 acquires classification condition information indicating each of a plurality of classification conditions for classifying the user U. The generation unit 35 generates, for each classification condition, a first content that corresponds to the classification condition and is used to classify a second content accessed by the user U, based on the classification condition information acquired by the acquisition unit 30. This allows the information processing device 1 to assist in the generation of the first content for classifying the second content.

[0251] The classification conditions are classification conditions based on values, and the generation unit 35 generates, as the first content, a content that is predicted to be accessed by a user U who has the values ​​indicated by the classification conditions. This allows the information processing device 1 to more appropriately support the generation of the first content for classifying the second content.

[0252] In addition, the generation unit 35 inputs input information including information instructing the generation of content predicted to be accessed by the user U who satisfies the classification condition to the generation AI, and causes the generation AI to generate the first content. This allows the information processing device 1 to more appropriately support the generation of the first content for classifying the second content.

[0253] The information processing device 1 also includes a collection unit 31 that collects a plurality of second contents and access information for each of the plurality of second contents, and an extraction unit 33 that extracts, from the plurality of second contents, second contents whose access information satisfies a predetermined condition, and a generation unit 35 causes the generation AI to generate a first content using information further including, as reference information, the second content extracted by the extraction unit 33 as input information. This allows the information processing device 1 to more appropriately support the generation of first contents for classifying the second contents.

[0254] Moreover, the information processing device 1 includes a receiving unit 32 that receives information indicating the classification conditions, and the acquiring unit 30 acquires, as classification condition information, the information indicating the classification conditions received by the receiving unit 32. This enables the information processing device 1 to more appropriately support the generation of the first content for classifying the second content.

[0255] Moreover, the information processing device 1 includes a collection unit 31 that collects a plurality of second contents, and a determination unit 34 that determines a plurality of classification conditions based on the plurality of second contents collected by the collection unit 31. This allows the information processing device 1 to more appropriately support the generation of first contents for classifying the second contents.

[0256] The second content is posted content posted on an online site, and the generating unit 35 generates, as the first content, posted content that is predicted to be accessed by a user U having the values ​​indicated by the classification conditions. This allows the information processing device 1 to more appropriately support the generation of the first content for classifying the second content.

[0257] Moreover, the online site is a Q&A site, and the posted content is a question. This enables the information processing device 1 to assist in generating a first content for classifying questions on the Q&A site.

[0258] Moreover, the online site is an EC site, and the posted content is product descriptions. This enables the information processing device 1 to assist in generating first content for classifying product descriptions on the EC site.

[0259] Moreover, the online site is a news site, and the posted content is a news article. This enables the information processing device 1 to assist in generating first content for classifying the news articles of the news site.

[0260] The information processing device 1 also includes a first classification unit 36 ​​that classifies second contents based on the first contents for each classification condition generated by the generation unit 35, and a second classification unit 37 that classifies users U based on access information of the users U to the second contents classified by the first classification unit 36. This allows the information processing device 1 to assist in classifying users U.

[0261] The above describes the embodiment of the present application in detail based on the drawings, but this is merely an example, and the present invention can be implemented in other forms that include various modifications and improvements based on the knowledge of those skilled in the art, including the forms described in the Disclosure of the Invention section.

[0262] Furthermore, the above-mentioned "section, module, unit" can be read as "means" or "circuit", etc. For example, an acquisition section can be read as an acquisition means or an acquisition circuit. [Explanation of symbols]

[0263] 1,4 Information processing device 2,3 Terminal Equipment 10. Communications Department 11 Storage section 12 Processing section 20 User information storage unit 21 Classification information storage section 22 First content storage unit 23 Second content storage unit 24 Classification result storage unit 30 Acquisition Department 31 Collection Department 32 Reception 33 Extraction part 34 Decision Section 35 Generation part 36 1st classification part 37 Second Classification Division 38 Third Classification Division 39 Providing Department 100 Information Processing Systems N Network

Claims

1. An acquisition unit that acquires classification condition information indicating each of a plurality of classification conditions for classifying users; a generation unit that generates, for each classification condition, first content according to the classification condition information acquired by the acquisition unit, the first content used to classify second content accessed by the user, the first content corresponding to the classification condition.

23. An information processing apparatus comprising:

2. The classification conditions are: It is a classification criteria based on values, The generation unit is Generate, as the first content, content that is predicted to be accessed by a user having the values ​​indicated by the classification conditions.

2. The information processing apparatus according to claim 1,

3. The generation unit is Inputting input information including information instructing generation of content predicted to be accessed by users who satisfy the classification condition to a generation AI, and causing the generation AI to generate the first content.

3. The information processing apparatus according to claim 2.

4. a collection unit that collects a plurality of the second contents and access information for each of the plurality of the second contents; an extracting unit that extracts a second content, the access information of which satisfies a predetermined condition, from the plurality of second contents; The generation unit is The generating AI generates the first content using information further including the second content extracted by the extracting unit as reference information as the input information.

4. The information processing apparatus according to claim 3.

5. a reception unit that receives information indicating the classification condition, The acquisition unit is Information indicating the classification conditions accepted by the accepting unit is acquired as the classification condition information.

5. The information processing apparatus according to claim 1, wherein the information processing apparatus is a computer.

6. A collection unit that collects a plurality of the second contents; a determination unit that determines a plurality of the classification conditions based on the plurality of the second contents collected by the collection unit.

4. The information processing apparatus according to claim 1, wherein the information processing apparatus is a computer.

7. The second content is is posted on an online site, The generation unit is Generate, as the first content, posted content that is predicted to be accessed by users having the values ​​indicated by the classification conditions.

5. The information processing device according to claim 2, wherein the information processing device is a computer.

8. The online site is It is a Q&A site, The posted content is: It is a question 8. The information processing apparatus according to claim 7,

9. The online site is It is an EC site, The posted content is: It is a product description 8. The information processing apparatus according to claim 7,

10. The online site is It is a news site, The posted content is: It is a news article 8. The information processing apparatus according to claim 7,

11. a first classification unit that classifies the second content based on the first content for each of the classification conditions generated by the generation unit; a second classification unit that classifies the users based on access information of the users to the second content classified by the first classification unit.

5. The information processing apparatus according to claim 1, wherein the information processing apparatus is a computer.

12. 1. A computer-implemented information processing method, comprising: An acquisition step of acquiring classification condition information indicating each of a plurality of classification conditions for classifying users; and generating, for each classification condition, first content that corresponds to the classification conditions and is used to classify second content accessed by the user, based on the classification condition information acquired by the acquiring step.

23. An information processing method comprising:

13. An acquisition step of acquiring classification condition information indicating each of a plurality of classification conditions for classifying users; a generating step of generating, for each classification condition, first content that corresponds to the classification condition and is used to classify second content accessed by the user, based on the classification condition information acquired by the acquiring step.

2. An information processing program comprising:

Images