Acquisition device and acquisition method

Abstract

A RAG system 20 according to the present disclosure comprises: a review reception unit 23 that receives review data including text describing a user's evaluation; a reference storage unit 27 that stores in advance score reference information expressing types of user emotion as text in association with a plurality of score levels for the evaluation; a generation unit 24 that generates input information which includes the review data received by the review reception unit 23 and the score reference information for each of the plurality of score levels read from the reference storage unit 27 and which requests the determination of a score for the review data; and a score acquisition unit 25 that acquires a score corresponding to the review data by inputting the input information to a generative AI model.

Description

Acquisition device and acquisition method

[0001] One aspect of the present disclosure relates to an acquisition device and an acquisition method.

[0002] In recent years, users have been registering reviews of products or services on networks such as websites as text. Other users can evaluate the products or services by referring to the registered reviews. A technology relating to a review-based evaluation method is described, for example, in Patent Document 1 listed below. In this technology, a system objectively evaluates a user's review and provides other users with the credibility of the user who wrote the review.

[0003] JP 2007-241983 A

[0004] Recently, there has been a trend for users to register reviews of many products or services, etc. Therefore, there is a need to automatically provide information on the level of user evaluations among multiple products or services.

[0005] Therefore, an object of the present disclosure is to provide an acquisition device and an acquisition method that can provide information regarding the level of user evaluations among a plurality of products or a plurality of services.

[0006] The acquisition device of the present disclosure includes a review acceptance unit that accepts review data including text describing a user's evaluation; a criteria storage unit that pre-stores score standard information that expresses the user's emotion type in text in association with multiple levels of scores related to the evaluation; a generation unit that generates input information that includes the review data accepted by the review acceptance unit and score standard information for each of the multiple levels of scores read from the criteria storage unit and requests determination of a score for the review data; and a score acquisition unit that acquires a score corresponding to the review data by inputting the input information into a generative AI model.

[0007] Alternatively, the acquisition method of the present disclosure is an acquisition method executed by an acquisition device, and includes: a review receiving step of receiving review data including text describing a user's evaluation; a criteria storage step of pre-storing score criteria information that expresses the user's emotion type in text, associated with multiple levels of scores related to the evaluation; a generation step of generating input information that includes the review data received by the review receiving step and score criteria information for each of the multiple levels of scores, and that requests determination of a score for the review data; and a score acquisition step of inputting the input information into a generative AI model to acquire a score corresponding to the review data.

[0008] According to one aspect of the present disclosure, information regarding the level of user evaluation among multiple products or services can be provided.

[0009] FIG. 1 is a block diagram showing a configuration of an acquisition system of the present disclosure. FIG. 2 is a diagram showing an example of the data configuration of review data acquired by the advance acquisition unit 21 of FIG. 1. FIG. 3 is a diagram showing an example of the data configuration of score criterion information stored in the criterion storage unit 27 of FIG. 1. FIG. 4 is a diagram showing an example of the data configuration of user attribute information stored in the attribute information storage unit 26 of FIG. 1. FIG. 5 is a diagram showing an example of the data configuration of score criterion information stored in the criterion storage unit 27 of FIG. 1. FIG. 6 is a diagram showing an example of the data configuration of review data accepted by the review acceptance unit 23 of FIG. 1. FIG. 7 is a diagram showing an example of response information acquired by the score acquisition unit 25 of FIG. 1. FIG. 8 is a diagram showing an example of the data configuration of output data generated by the score acquisition unit 25 of FIG. 1. FIG. 9 is a flowchart showing the procedure of score criterion information generation processing by the RAG system 20. FIG. 10 is a flowchart showing the procedure of score acquisition processing by the RAG system 20. FIG. 11 is a diagram showing an example of the hardware configuration of the RAG system 20 according to an embodiment of the present disclosure.

[0010] The present disclosure will be described with reference to the accompanying drawings. Whenever possible, the same parts are designated by the same reference numerals and redundant description will be omitted.

[0011] Fig. 1 is a diagram showing the device configuration of an acquisition system according to this embodiment. The acquisition system shown in Fig. 1 includes a database server device 10A, a terminal 10B, a Retrieval-Augmented Generation (RAG) system 20, and a server device 30, which are configured to be able to communicate with each other via a network including a wireless communication network and a fixed communication network. The RAG system 20 constitutes an acquisition device that acquires a score for review data received from the terminal 10B.

[0012] The terminal 10B is a device used by a user who wishes to obtain a score automatically generated based on review data using an interactive AI model. The terminal 10B may be, for example, a personal computer, a smartphone, a tablet terminal, a feature phone, a server device, a game console, or the like. Note that while only two terminals 10B are illustrated in FIG. 1 , the obtaining system may include any number of terminals 10B greater than or equal to two.

[0013] The database server device 10A is a device that accumulates review data including text in which each of multiple users, including the user of the terminal 10B, has previously written evaluations of a specific product or service. Each piece of review data accumulated by the database server device 10A is also accompanied by a score that each user manually sets by quantifying the evaluation level on a multiple-level scale. The database server device 10A is a device operated, for example, by a business that provides a product or service, a business that mediates the provision of a product or service, or a business that provides information about a product or service, and accumulates review data received from users who have received the product or service.

[0014] The server device 30 is a device that enables the provision of content using a generative AI model. The generative AI model is a model that can generate content in response to a prompt including input information, according to any one or a combination of the instructions, context, question, and output format indicated by the prompt, and return the content as response information. The prompt can also include input information, in which case the generative AI model generates response information targeted at the input information. The generative AI model may be, for example, an interactive AI model that includes a large language model (LLM) and a user interface (UI) for interacting with a user, and enables text or voice chat with the user.

[0015] The generative AI model used in this embodiment responds to input of a textual prompt with a textual response. Examples of such generative AI models include ChatGPT, GPT®-3.5, GPT-3.5 Turbo, GPT-4.0, GPT-4.0 Turbo, Azure OpenAI Service, Tsuzumi, GPT-4V, and PaLM2.

[0016] In this embodiment, the server device 30 is capable of providing a content provision function using an interactive AI model 31. This interactive AI model 31 may be stored in the server device 30, or may be stored in another device connected to the server device 30 via a network and configured to enable information exchange with a user via the server device 30. The interactive AI model 31 may also be stored in the terminal 10B. Note that while only one server device 30 is illustrated in FIG. 1, the acquisition system may include multiple server devices 30. Also, although the above describes an example of a large-scale language model, other AI models may also be used. Also, an AI model selected from multiple types of AI models may be used in the acquisition system.

[0017] RAG system 20 is configured to include, as functional components, a pre-acquisition unit 21, a criteria generation unit 22, a review acceptance unit 23, a generation unit 24, a score acquisition unit 25, an attribute information storage unit 26, and a criteria storage unit 27. RAG system 20 relays prompts (input information) generated based on data acquired from terminal 10B or database server device 10A to server device 30, and acquires response information from server device 30 in response to the prompts. The functions of each functional unit of RAG system 20 will be described in detail below.

[0018] The pre-acquisition unit 21 acquires from the database server device 10A multiple review data sets for multiple products or services previously received and accumulated from multiple users, including multiple levels of scores previously assigned by the users to each of the multiple review data sets. FIG. 2 shows an example of the data structure of the review data acquired by the pre-acquisition unit 21. Each review data set includes a user ID, which is user identification information identifying the user who created the review data, a title indicating the subject of the user's evaluation, a review, which is a sentence (text) describing the user's evaluation, and a numerical score indicating the user's evaluation level for the product or service with the title, all associated with each other. For example, the review "A violent beauty filled with love. The relationship was moving, and the final scene brought me to tears" and a score of "9" are associated with the user ID "AAA" and the title "Movie A." The pre-acquisition unit 21 passes the acquired multiple review data sets to the criteria generation unit 22.

[0019] The criterion generation unit 22 extracts a portion of the review data from the multiple review data acquired by the advance acquisition unit 21 in accordance with the extraction conditions, generates score criterion information for each of multiple levels of scores based on the portion of the review data, and stores the score criterion information in the criterion storage unit 27 in association with the multiple levels of scores.

[0020] That is, when the extraction conditions are set to extraction on a per-user basis, the reference generation unit 22 extracts a portion of review data for each user ID, which is user identification information. Then, from the extracted portion of review data, the reference generation unit 22 identifies reviews corresponding to scores around one level, extracts text from the identified reviews that expresses the user's emotional type, and generates score reference information corresponding to one level of score, including text that is identical or similar to the extracted text. Furthermore, the reference generation unit 22 performs the above operation for scores at multiple levels, thereby generating score reference information corresponding to multiple levels.

[0021] For example, the criteria generation unit 22 can generate score criteria information using the interactive AI model 31. In this case, the criteria generation unit 22 generates a first prompt that includes reviews and scores for each record in a portion of the review data extracted for each user ID and requests the generation of score criteria information for each of multiple levels of scores corresponding to the user ID. A prompt is information indicating an instruction or question entered by a user in an interactive system such as a dialogue with an AI model or a command line interface (CLI). The prompt generated by the criteria generation unit 22 expresses, in text, for example, commands to be executed by the interactive AI model, tasks to be executed by the interactive AI model, background / context to be considered by the interactive AI model (e.g., roles, conditions), questions to be answered by the interactive AI model, and the output format of response information from the interactive AI model. Furthermore, input information to be the target of commands / tasks to be executed by the interactive AI model may be added to the prompt. Examples of such input information include data files such as text data, image data, audio data, video data, and still image data.

[0022] Furthermore, the criterion generating unit 22 inputs the generated first prompt to the interactive AI model 31 and acquires score criterion information for each of the multiple levels of scores from the interactive AI model 31. The criterion generating unit 22 repeats the generation of score criterion information for each of the multiple levels of scores for multiple user IDs, and stores the generated score criterion information for each of the multiple levels of scores in the criterion storage unit 27 in association with the user ID.

[0023] An example of a first prompt generated by the criteria generating unit 22 using a portion of the review data corresponding to the user ID "AAA" extracted from the review data shown in FIG. 2 is as follows: "Create a scoring scale from the following reviews and scores. # Conditions - Find axes related to scores from reviews - Determine scoring criteria based on the axes # Input Review Score A violent beauty filled with love. You'll be moved by the relationship and teary-eyed by the final scene. 9 A story of love and courage. A masterpiece filled with laughter, tears, and all kinds of emotions. 9 More than the epic voyage of love and despair, you'll be captivated by the story of Jack and Rose. 8.5 A fun film with unique humor and a thrilling plot. 6 Moved by the heartfelt acting and encouraging story. 5 A gem of entertainment with an exquisite blend of laughter and thrills. 4 A new paradigm in film. Amazed by the original storytelling. 3 The pinnacle of horror. Eerie music and heart-stopping acting. Extremely scary! 2 A moving essay that resonates with hope and indomitable spirit. 2 A new height for superhero movies. The dark and spectacular world of X-Man. 1 # Output Score 1 Axis: Score 2 axis: Score 3 axis: Score 4 axis: Score 5 axis: Score 6 axis: Score 7 axis: Score 9 axis: Score 10 axis:

[0024] Examples of score criterion information for each of the multiple levels of scores acquired by the criterion generation unit 22 in response to the input of the first prompt are as follows: "Score 1 axis: the grandeur and dark worldview of a superhero movie Score 2 axis: fear and eeriness Score 3 axis: originality and amazing storytelling Score 4 axis: an exquisite balance of laughter and thrills Score 5 axis: a touching story and inspiring acting Score 6 axis: humor and thrilling developments Score 7 axis: entertaining appeal Score 9 axis: relationships that evoke affection and emotion Score 10 axis: the emotion of a masterpiece filled with love and courage" Figure 3 shows an example of the data configuration of score criterion information stored in the criterion storage unit 27 in accordance with the above example.

[0025] On the other hand, when the extraction condition is set to extraction by user attribute unit, the criterion generation unit 22 first extracts multiple user IDs corresponding to the attributes of one user by referring to the attribute information storage unit 26, which stores user attribute information indicating attributes of multiple users. Furthermore, the criterion generation unit 22 extracts a portion of review data corresponding to the extracted multiple user IDs. Then, the criterion generation unit 22 generates score criterion information corresponding to multiple levels based on the extracted portion of review data in a procedure similar to the procedure described above. The criterion generation unit 22 repeatedly generates score criterion information for each score at multiple levels for multiple user attributes, and stores the generated score criterion information for each score at multiple levels in the criterion storage unit 27 in association with the user attributes.

[0026] FIG. 4 shows an example of the data configuration of user attribute information stored in the attribute information storage unit 26. Thus, the attribute information storage unit 26 stores data on residence, age, and attributes 1 to 5, which are attribute information indicating the user's attributes, in association with a user ID that identifies the user. For example, the following data is associated with the user ID "AAA": residence "Japan," age "31," attribute 1 "2 days / week" indicating the frequency of remote work, attribute 2 "3 days / week" indicating the frequency of drinking, attribute 3 "detached house" indicating the type of residence, attribute 4 "family" indicating the household composition, and attribute 5 "yes" indicating whether or not the user keeps pets. These pieces of attribute information may be set, for example, based on response data to a questionnaire that each user previously sent using the terminal 10B or the like. For example, by referring to the user attribute information having the above data configuration, the criteria generation unit 22 extracts user IDs "AAA" and "BBB" having a place of residence "Japan" as attribute information, and stores score criteria information generated using a portion of review data corresponding to the extracted user IDs "AAA" and "BBB" in association with the user attribute of place of residence "Japan" in the criteria storage unit 27. Figure 5 shows an example of score criteria information stored in association with the user attribute of place of residence "Japan."

[0027] Furthermore, the criterion generating unit 22 may extract user IDs "BBB" and "CCC" having ages "45" and "46" included in the age group of the 40s as attribute information, and may store score criterion information generated using a portion of review data corresponding to the extracted user IDs "BBB" and "CCC" in association with the age group "40s," which is a group of user attributes, in the criterion storage unit 27. Furthermore, the criterion generating unit 22 may generate score criterion information corresponding to a combination of multiple attributes (for example, a combination of place of residence "Japan" and age group "40s").

[0028] Here, when generating score criteria information for quantifiable user attributes, the criteria generation unit 22 may divide user IDs into groups using a clustering method such as k-means and generate score criteria information for each user group. Furthermore, when generating score criteria information for qualitative user attributes, the criteria generation unit 22 may also divide user IDs into groups and generate score criteria information for each user group. For example, in the example of user attribute information shown in FIG. 4 , when grouping based on attributes 3 and 4, the criteria generation unit 22 groups the user ID "AAA," the user ID "BBB," and the user ID "CCC" into separate groups. When grouping based on attributes 3 and 5, the criteria generation unit 22 groups the user ID "AAA" and the two user IDs "BBB" and "CCC" into separate groups.

[0029] The criterion generating unit 22 may also refer to scores previously assigned by multiple users to multiple products or services and generate score criterion information based on a portion of review data corresponding to multiple users who share similar values. That is, the criterion generating unit 22 may calculate a correlation coefficient indicating the degree of similarity between scores assigned to users, extract multiple user IDs for a specific user whose correlation coefficient is equal to or greater than a certain value, and generate score criterion information based on a portion of review data corresponding to the extracted multiple user IDs. Alternatively, the criterion generating unit 22 may group users using a clustering method such as k-means based on a correlation function and generate score criterion information for each user group.

[0030] Referring back to FIG. 1 , the review receiving unit 23 receives one or more review data items for which a score is to be obtained from the terminal 10B. FIG. 6 shows an example of the data structure of the review data received by the review receiving unit 23. As described above, the review data includes a title that identifies the product or service to be rated by the user, a review that is a statement describing the user's evaluation, and a pre-score that indicates the level of the evaluation that the user manually set in advance. For example, in one piece of review data, the title "Movie K" is associated with the review "The dark and sad story leaves a vivid impression" and the pre-score of "9."

[0031] The generation unit 24 generates a second prompt requesting determination of a score for the review data, the second prompt including the review data received by the review receiving unit 23 and score criterion information for each of multiple score levels read from the criterion storage unit 27 according to preset extraction conditions. For example, when the extraction conditions are set to extraction on a user-by-user basis, the generation unit 24 includes, in the second prompt, score criterion information stored in association with a user ID indicating the user who created the review data. On the other hand, when the extraction conditions are set to extraction on a user attribute basis, the generation unit 24 includes, in the second prompt, score criterion information stored in association with the attribute of the user who created the review data. Furthermore, when the extraction conditions are set to extraction on a user group basis, the generation unit 24 includes, in the second prompt, score criterion information stored in association with a group including the user who created the review data.

[0032] An example of a second prompt generated by the generating unit 24 using the example of score criteria information shown in FIG. 3 and the example of review data shown in FIG. 6 is as follows: Please rate your review according to the following criteria. # Rating criteria Score 1 Axis: The grandeur and dark worldview of a superhero movie Score 2 Axis: Fear and creepiness Score 3 Axis: Originality and astonishing storytelling Score 4 Axis: An exquisite balance of laughter and thrills Score 5 Axis: A heartwarming story and inspiring acting Score 6 Axis: Humor and thrilling developments Score 7 Axis: The appeal of entertainment Score 9 Axis: A relationship that evokes love and emotion Score 10 Axis: The moving masterpiece filled with love and courage # Review The dark and heartbreaking story leaves a vivid impression You will be captivated by the vibrant and moving story A romantic drama unfolding on a spectacular stage will touch your heart A stylish and humorous masterpiece of a gangster movie A heartwarming film that will move you with a story full of growth and lessons A fun horror comedy that will surprise and make you laugh A unique darkness and fascinating characters make this a brilliant crime movie to be enjoyed A chilling psychological horror film with chilling tension. A moving drama with a poignant journey to hope and freedom. An overwhelming X-Man film with a dark, intricate story and powerful action. # Output Review, Score, Reason

[0033] The score acquiring unit 25 inputs the second prompt generated by the generating unit 24 to the interactive AI model 31, and acquires response information including a score automatically assigned corresponding to the review data from the interactive AI model 31. Figure 7 shows an example of response information acquired by the score acquiring unit 25 using the above example of the second prompt. In this way, the score acquired corresponding to the title indicating the target to which the score is assigned and data on the reason indicating the basis for the score evaluation are acquired.

[0034] Furthermore, the score acquisition unit 25 has a function of outputting the score acquired from the interactive AI model 31 as a final score when the difference between the score acquired from the interactive AI model 31 and the preliminary score accepted by the review acceptance unit 23 exceeds a preset difference (e.g., "5"). For example, the score acquisition unit 25 outputs data in which the preliminary score in the review data accepted by the review acceptance unit 23 is rewritten as a final score. Figure 8 shows an example of output data in which the score acquisition unit 25 has rewritten the review data shown in Figure 6 into a final score. The output data by the score acquisition unit 25 may be output by transmitting it to an external device such as terminal 10B, or by storing it in a data storage unit within the RAG system 20 so that it can be referenced from outside.

[0035] The procedure for score acquisition processing by the RAG system 20 configured as described above, i.e., the flow of the acquisition method according to this embodiment, will be described below. Fig. 9 is a flowchart showing the procedure for score reference information generation processing by the RAG system 20, and Fig. 10 is a flowchart showing the procedure for score acquisition processing by the RAG system 20.

[0036] First, a score criterion information generation process is executed prior to the score acquisition process for review data created by a user. Referring to FIG. 9 , when the score criterion information generation process is initiated, the pre-acquisition unit 21 of the RAG system 20 first acquires multiple pieces of review data created in the past by multiple users from the database server device 10A (step S01). Then, the criterion generation unit 22 extracts a portion of the review data from the multiple pieces of review data in accordance with the extraction conditions (step S02). Then, the criterion generation unit 22 generates score criterion information for each of multiple score levels based on the portion of the review data (step S03). Finally, the criterion generation unit 22 stores the generated score criterion information in the criterion storage unit 27 (step S04). The processes of steps S02 to S04 are repeated multiple times according to the set extraction conditions, and score criterion information generated for each user ID, user attribute, or user group is stored in the criterion storage unit 27.

[0037] 10 , when the score acquisition process is started after the score criterion information generation process is executed, the review receiving unit 23 first receives review data for which a score is to be acquired (step S101). Next, the generation unit 24 reads score criterion information for each of multiple score levels from the criterion storage unit 27 according to the extraction conditions (step S102). For example, if the extraction conditions are set to extraction on a per-user basis, the score criterion information associated with the user ID indicating the user who created the review data is read. Then, the generation unit 24 generates a second prompt including the review data and the read score criterion information (step S103).

[0038] Thereafter, the score acquiring unit 25 inputs the second prompt to the interactive AI model 31 to acquire response information (step S104). Next, the score acquiring unit 25 generates output data in which the preliminary score included in the review data is rewritten with the final score included in the response information (step S105). Finally, the score acquiring unit 25 transmits the generated output data to the terminal 10B (step S106).

[0039] Next, the effects of the acquisition device of the present disclosure will be described. According to the RAG system 20 of the present disclosure, review data including text describing an evaluation is received from the terminal 10B. A second prompt is generated that includes the review data and score criterion information for each of multiple scores read from the criterion storage unit 27 and requests a score for the review data. The generated second prompt is then input to the interactive AI model 31, whereby a score corresponding to the review data is acquired. As a result, a score determined by comparing the preset score criterion information with the text included in the review data can be acquired, and the relative evaluation level of a product or service that is the subject of the review data compared to other products or services can be acquired. As a result, information regarding the user evaluation levels among multiple products or services can be provided.

[0040] Here, the RAG system 20 of the present disclosure further includes a pre-acquisition unit 21 that acquires multiple pieces of review data and multiple levels of scores previously assigned to each of the multiple pieces of review data, and a criterion generation unit 22 that stores text extracted from the multiple pieces of review data as score criterion information in the criterion storage unit 27 in association with each of the multiple levels of scores. This makes it possible to store criterion information for each score in the criterion storage unit 27 using multiple pieces of review data previously created by users and the scores previously assigned by users to the multiple pieces of review data. As a result, score criterion information associated with scores actually assigned by users can be registered, making it possible to provide information on highly reliable evaluation levels.

[0041] In the RAG system 20 of the present disclosure, the criteria storage unit 27 pre-stores score criteria information for each of multiple user identification information, and the generation unit 24 includes, in the second prompt, score criteria information for each of multiple score levels associated with the user identification information indicating the user. In this case, a score determined using score criteria information corresponding to the user who created the review data from among the score criteria information pre-set for each of multiple users can be obtained. As a result, an evaluation level that matches the user's evaluation criteria can be obtained.

[0042] The RAG system 20 of the present disclosure further includes an attribute information storage unit 26 that stores attribute information indicating user attributes for each of multiple user identification information, a criterion storage unit 27 that pre-stores score criterion information for each of the multiple user attributes, and a generation unit 24 that extracts attribute information corresponding to the user identification information indicating the user from the attribute information storage unit 26 and includes score criterion information for each of multiple score levels corresponding to the attributes indicated by the extracted attribute information in the second prompt. This makes it possible to obtain a score determined using score criterion information corresponding to the attributes of the user who created the review data from the score criterion information pre-set for each of the multiple user attributes. As a result, it is possible to obtain an evaluation level that matches the evaluation criterion corresponding to the user's attributes.

[0043] In the RAG system 20 of the present disclosure, the review receiving unit 23 further receives a pre-score, which is a score previously assigned to the review data. If the difference between the acquired score and the pre-score exceeds a predetermined difference, the score acquisition unit 25 outputs the score as the final score assigned to the review data. With this configuration, when review data previously assigned a pre-score is received, a score that is evaluated as inappropriate in light of the user's evaluation criteria can be corrected and output as the final score. As a result, a level of evaluation corrected in light of the user's evaluation criteria can be acquired.

[0044] The acquisition device and acquisition method of the present disclosure have the following configuration.

[0045] [1] An acquisition device comprising: a review receiving unit that receives review data including text describing a user's evaluation; a criteria storage unit that stores in advance score criteria information that expresses a type of user emotion in text in association with multiple levels of scores related to the evaluation; a generation unit that generates input information that includes the review data received by the review receiving unit and the score criteria information for each of the multiple levels of scores read from the criteria storage unit, and requests determination of a score for the review data; and an acquisition unit that inputs the input information to a generative AI model to acquire a score corresponding to the review data.

[0046] [2] The acquisition device according to [1] above, further comprising: a score receiving unit that receives a plurality of the review data and a plurality of levels of scores previously assigned to each of the plurality of review data; and a criterion generating unit that stores text extracted from the plurality of review data in the criterion storage unit as the score criterion information, in association with each of the plurality of levels of scores.

[0047] [3] The acquisition device according to [1] or [2] above, wherein the standard storage unit pre-stores the score standard information for each of a plurality of pieces of user identification information, and the generation unit includes, in the input information, the score standard information for each of the plurality of levels of scores associated with the user identification information indicating the user.

[0048] [4] The acquisition device according to [1] or [2] above, further comprising an attribute storage unit that stores attribute information indicating user attributes for each of a plurality of pieces of user identification information, wherein the standard storage unit pre-stores the score standard information for each of the plurality of pieces of user identification information, and the generation unit extracts from the attribute storage unit a plurality of pieces of user identification information having the attribute information indicating the same attributes as the attributes of the user, and includes in the input information the score standard information for each of the plurality of levels of scores having the extracted plurality of pieces of user identification information.

[0049] [5] The acquisition device according to any one of [1] to [4] above, wherein the review receiving unit further receives a pre-score that is a score assigned to the review data in advance, and the acquisition unit outputs the score as a final score assigned to the review data when a difference between the acquired score and the pre-score exceeds a preset difference.

[0050] [6] An acquisition method executed by an acquisition device, comprising: a review receiving step of receiving review data including text describing a user's evaluation; a criteria storing step of pre-storing score criteria information, in which the type of user emotion is expressed in text, in association with a plurality of levels of scores related to the evaluation; a generation step of generating input information including the review data received by the review receiving step and the score criteria information for each of the plurality of levels of scores, and requesting determination of a score for the review data; and an acquisition step of inputting the input information into a generative AI model to acquire a score corresponding to the review data.

[0051] The block diagrams used to explain the above embodiments show functional blocks. These functional blocks (components) are realized by any combination of hardware and / or software. Furthermore, the method for realizing each functional block is not particularly limited. That is, each functional block may be realized using a single device that is physically or logically coupled, or may be realized using two or more physically or logically separated devices that are connected directly or indirectly (e.g., via wire, wirelessly, etc.) and these multiple devices. The functional block may also be realized by combining the single device or multiple devices with software.

[0052] Functions include, but are not limited to, judgment, determination, assessment, calculation, computation, processing, derivation, investigation, search, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, consideration, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating, mapping, and assignment. For example, a functional block (component) that performs transmission is called a transmitting unit or transmitter. As mentioned above, there are no particular limitations on how these functions are implemented.

[0053] For example, the RAG system 20 constituting the acquisition device according to an embodiment of the present disclosure may function as a computer that performs processing of the acquisition method of the present disclosure. FIG. 11 is a diagram illustrating an example of the hardware configuration of the RAG system 20 according to an embodiment of the present disclosure. The RAG system 20 described above may be physically configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input device 1005, an output device 1006, a bus 1007, and the like. The RAG system 20 may be configured as a computer device including at least one processor such as a CPU or GPU, or may be configured as a computer device including multiple processors or may include multiple computer devices. The terminal 10B, the database server device 10A, and the server device 30 may also have a similar hardware configuration.

[0054] In the following description, the term "apparatus" can be interpreted as a circuit, a device, a unit, etc. The hardware configuration of the RAG system 20 may be configured to include one or more of the apparatuses shown in the figure, or may be configured to exclude some of the apparatuses.

[0055] Each function in the RAG system 20 is realized by loading specified software (programs) onto hardware such as the processor 1001 and memory 1002, causing the processor 1001 to perform calculations, control communication via the communication device 1004, and control at least one of reading and writing data in the memory 1002 and storage 1003.

[0056] The processor 1001 controls the entire computer by running, for example, an operating system. The processor 1001 may be configured by a central processing unit (CPU) including an interface with peripheral devices, a control device, an arithmetic unit, a register, etc. For example, the above-mentioned pre-acquisition unit 21, criteria generation unit 22, review acceptance unit 23, generation unit 24, score acquisition unit 25, etc. may be realized by the processor 1001.

[0057] The processor 1001 also reads programs (program codes), software modules, data, etc. from at least one of the storage 1003 and the communication device 1004 into the memory 1002 and executes various processes in accordance with the programs. The programs used include programs that cause a computer to execute at least some of the operations described in the above-described embodiments. For example, the pre-acquisition unit 21, the criteria generation unit 22, the review acceptance unit 23, the generation unit 24, and the score acquisition unit 25 may be implemented by a control program stored in the memory 1002 and running on the processor 1001, and similar implementations may be used for other functional blocks. While the above-described various processes have been described as being executed by a single processor 1001, they may also be executed simultaneously or sequentially by two or more processors 1001. The processor 1001 may be implemented by one or more chips. The programs may also be transmitted from a network via a telecommunications line.

[0058] The memory 1002 is a computer-readable recording medium and may be configured by, for example, at least one of a read-only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a random access memory (RAM), etc. The memory 1002 may also be called a register, a cache, a main memory (primary storage device), etc. The memory 1002 can store executable programs (program codes), software modules, etc. for implementing an acquisition method according to an embodiment of the present disclosure.

[0059] Storage 1003 is a computer-readable recording medium, and may be composed of at least one of, for example, an optical disk such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disk, a digital versatile disk, a Blu-ray (registered trademark) disk), a smart card, a flash memory (e.g., a card, a stick, a key drive), a floppy (registered trademark) disk, a magnetic strip, etc. Storage 1003 may also be referred to as an auxiliary storage device. The above-mentioned storage medium may be, for example, a database, a server, or other appropriate medium including at least one of memory 1002 and storage 1003.

[0060] The communication device 1004 is hardware (transmission / reception device) for communicating between computers via at least one of a wired network and a wireless network, and is also referred to as, for example, a network device, a network controller, a network card, or a communication module. The communication device 1004 may be configured to include a high-frequency switch, a duplexer, a filter, a frequency synthesizer, etc. to realize at least one of frequency division duplex (FDD) and time division duplex (TDD). For example, the above-mentioned advance acquisition unit 21, review acceptance unit 23, score acquisition unit 25, etc. may be realized by the communication device 1004.

[0061] The input device 1005 is an input device (e.g., a keyboard, a mouse, a microphone, a switch, a button, a sensor, etc.) that accepts input from the outside. The output device 1006 is an output device (e.g., a display, a speaker, an LED lamp, etc.) that outputs to the outside. Note that the input device 1005 and the output device 1006 may be integrated into one device (e.g., a touch panel).

[0062] Furthermore, each device, such as the processor 1001 and the memory 1002, is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or may be configured using different buses between each device.

[0063] Furthermore, the RAG system 20 may be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a field programmable gate array (FPGA), and some or all of the functional blocks may be realized by the hardware. For example, the processor 1001 may be implemented using at least one of these pieces of hardware.

[0064] The notification of information is not limited to the aspects / embodiments described in the present disclosure and may be performed using other methods. For example, the notification of information may be performed by physical layer signaling (e.g., Downlink Control Information (DCI) and Uplink Control Information (UCI)), higher layer signaling (e.g., Radio Resource Control (RRC) signaling, Medium Access Control (MAC) signaling, broadcast information (Master Information Block (MIB) and System Information Block (SIB))), other signals, or a combination thereof. Furthermore, the RRC signaling may be referred to as an RRC message, and may be, for example, an RRC Connection Setup message, an RRC Connection Reconfiguration message, or the like.

[0065] The order of the procedures, sequences, flowcharts, etc. of each aspect / embodiment described in this disclosure may be changed unless it is consistent. For example, the methods described in this disclosure present elements of various steps using an example order, and are not limited to the particular order presented.

[0066] Input and output information may be stored in a specific location (for example, memory) or may be managed using a management table. Input and output information may be overwritten, updated, or added to. Output information may be deleted. Input information may be sent to another device.

[0067] The determination may be made based on a value represented by one bit (0 or 1), a Boolean value (true or false), or a numerical comparison (e.g., comparison with a predetermined value).

[0068] The aspects / embodiments described in this disclosure may be used alone, in combination, or switched depending on the implementation. Notification of predetermined information (e.g., notification that "X is true") is not limited to explicit notification, but may be implicit (e.g., not notifying the predetermined information).

[0069] Although the present disclosure has been described in detail above, it is clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the spirit and scope of the present disclosure as defined by the claims. Therefore, the description of the present disclosure is intended to be illustrative and does not have any limiting meaning on the present disclosure.

[0070] The acquisition system is not limited to the configuration shown in FIG. 1 , and may have a configuration in which server device 30 is implemented in terminal 10B. This configuration can be realized by installing an application that executes the functions of server device 30 in terminal 10B. Also, the RAG system 20 may be implemented in terminal 10B. This configuration can be realized by installing an application that executes the functions of RAG system 20 in terminal 10B. Also, while FIG. 1 shows an example in which attribute information storage unit 26 and reference storage unit 27 are implemented outside terminal 10B (for example, on a network), attribute information storage unit 26 and reference storage unit 27 may be implemented in terminal 10B.

[0071] Software shall be construed broadly to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, threads of execution, procedures, functions, etc., whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise.

[0072] Software, instructions, information, etc. may also be transmitted or received over a transmission medium. For example, if software is transmitted from a website, server, or other remote source using wired technologies (such as coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL)), and / or wireless technologies (such as infrared, microwave), then these wired and / or wireless technologies are included within the definition of transmission media.

[0073] The information, signals, etc. described in this disclosure may be represented using any of a variety of different technologies. For example, data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof.

[0074] Note that terms described in this disclosure and terms necessary for understanding this disclosure may be replaced with terms having the same or similar meanings. For example, at least one of a channel and a symbol may be a signal (signaling). Furthermore, a signal may be a message. Furthermore, a component carrier (CC) may be called a carrier frequency, a cell, a frequency carrier, etc.

[0075] Furthermore, the information, parameters, etc. described in the present disclosure may be expressed using absolute values, may be expressed using relative values ​​from a predetermined value, or may be expressed using other corresponding information. For example, a radio resource may be indicated by an index.

[0076] The names used for the above-described parameters are not intended to be limiting in any way. Furthermore, the mathematical expressions using these parameters may differ from those explicitly disclosed in this disclosure. The various channels (e.g., PUCCH, PDCCH, etc.) and information elements may be identified by any suitable names, and therefore the various names assigned to these various channels and information elements are not intended to be limiting in any way.

[0077] In this disclosure, the terms "Mobile Station (MS)," "user terminal," "User Equipment (UE)," "terminal," etc. may be used interchangeably.

[0078] A mobile station may also be referred to by those skilled in the art as a subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, access terminal, mobile terminal, wireless terminal, remote terminal, handset, user agent, mobile client, client, or some other suitable terminology.

[0079] As used in this disclosure, the terms "determining" and "determining" may encompass a wide variety of actions. "Determining" and "determining" may include, for example, judging, calculating, computing, processing, deriving, investigating, looking up, searching, inquiring (e.g., searching in a table, database, or other data structure), ascertaining, and the like. "Determining" and "determining" may also include receiving (e.g., receiving information), transmitting (e.g., sending information), input, output, accessing (e.g., accessing data in memory), and the like. Furthermore, "judgment" and "decision" can include regarding resolving, selecting, choosing, establishing, comparing, etc. as having been "judged" or "decided." In other words, "judgment" and "decision" can include regarding some action as having been "judged" or "decided." Furthermore, "judgment (decision)" can be interpreted as "assuming," "expecting," "considering," etc.

[0080] The terms "connected," "coupled," or any variation thereof, refer to any direct or indirect connection or coupling between two or more elements, and may include the presence of one or more intermediate elements between two elements that are "connected" or "coupled" to each other. The coupling or connection between elements may be physical, logical, or a combination thereof. For example, "connected" may be read as "access." As used in this disclosure, two elements may be considered to be "connected" or "coupled" to each other using one or more wires, cables, and / or printed electrical connections, as well as electromagnetic energy having wavelengths in the radio frequency range, microwave range, and optical (both visible and invisible) range, as some non-limiting and non-exhaustive examples.

[0081] As used in this disclosure, the phrase "based on" does not mean "based only on," unless expressly stated otherwise. In other words, the phrase "based on" means both "based only on" and "based at least on."

[0082] As used in this disclosure, any reference to an element using a designation such as "first," "second," etc. does not generally limit the quantity or order of those elements. These designations may be used in this disclosure as a convenient method of distinguishing between two or more elements. Thus, a reference to a first and a second element does not imply that only two elements may be employed or that the first element must in some way precede the second element.

[0083] When the terms "include," "including," and variations thereof are used in this disclosure, these terms are intended to be inclusive, similar to the term "comprising." Furthermore, when the term "or" is used in this disclosure, it is not intended to be an exclusive or.

[0084] In this disclosure, where articles are added by translation, such as a, an, and the in English, the disclosure may include that the nouns following these articles are in the plural form.

[0085] In the present disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "coupled" may also be interpreted in the same way as "different."

[0086] 10B...terminal, 20...RAG system, 30...server device, 31...interactive AI model, 21...pre-acquisition unit, 22...criteria generation unit, 23...review acceptance unit, 24...generation unit, 25...score acquisition unit, 26...attribute information storage unit, 27...criteria storage unit.

Claims

1. An acquisition device comprising: a review acceptance unit that accepts review data including text describing a user's evaluation; a criteria storage unit that pre-stores score criteria information that expresses the user's emotional types in text, associated with multiple levels of scores related to the evaluation; a generation unit that generates input information that includes the review data accepted by the review acceptance unit and the score criteria information for each of the multiple levels of scores read from the criteria storage unit, and requests the determination of a score for the review data; and a score acquisition unit that acquires a score corresponding to the review data by inputting the input information into a generative AI model.

2. The acquisition device of claim 1, further comprising: a pre-acquisition unit that acquires a plurality of the review data and a plurality of levels of scores that have been assigned in advance to each of the plurality of review data; and a criterion generation unit that stores text extracted from the plurality of review data in the criterion storage unit as the score criterion information, in association with each of the plurality of levels of scores.

3. The acquisition device described in claim 1, wherein the standard storage unit pre-stores the score standard information for each of multiple user identification information, and the generation unit includes the score standard information for each of the multiple levels of scores associated with the user identification information indicating the user in the input information.

4. The acquisition device of claim 1, further comprising an attribute storage unit that stores attribute information indicating user attributes for each of a plurality of user identification information, wherein the standard storage unit pre-stores the score standard information for each of a plurality of user attributes, and the generation unit extracts the attribute information corresponding to the user identification information indicating the user from the attribute storage unit, and includes the score standard information for each of the plurality of levels of scores corresponding to the attributes indicated by the extracted attribute information in the input information.

5. The acquisition device described in claim 1, wherein the review receiving unit further receives a pre-score, which is a score previously assigned to the review data, and the score acquisition unit outputs the score as the final score assigned to the review data when the difference between the acquired score and the pre-score exceeds a predetermined difference.

6. An acquisition method executed by an acquisition device, comprising: a review receiving step of receiving review data including text describing a user's evaluation; a criteria storage step of pre-storing score criteria information that expresses the user's emotion type in text, in association with multiple levels of scores related to the evaluation; a generation step of generating input information that includes the review data received by the review receiving step and the score criteria information for each of the multiple levels of scores, and that requests determination of a score for the review data; and a score acquisition step of inputting the input information into a generative AI model to acquire a score corresponding to the review data.

Images