WO2021200502A1

WO2021200502A1 - Information processing device and information processing method

Info

Publication number: WO2021200502A1
Application number: PCT/JP2021/012365
Authority: WO
Inventors: 眞大山本; 加奈西川; 知香明賀; 康治浅野; 浩明小川; 典子戸塚; 高橋　晃; ミヒャエルヘンチェル; 智恵山田; 匡伸中村
Original assignee: ソニーグループ株式会社
Priority date: 2020-03-31
Filing date: 2021-03-24
Publication date: 2021-10-07
Also published as: JP2021162997A

Abstract

An information processing device according to the present invention comprises: an acquisition unit for acquiring text information that indicates a use case designated by a user; and an estimation unit for estimating, on the basis of the text information, a recommended value for a corresponding specification that corresponds to the use case among the specifications of an intended transaction category corresponding to the use case.

Description

Information processing device and information processing method

This disclosure relates to an information processing device and an information processing method.

Sales via networks such as the Internet have become popular. For example, electronic commerce (EC: Electronic Commerce) provides technology for selling transaction targets such as goods and services to users. For example, when a user purchases a product via the Internet, the dialogue with the user is controlled in order to provide an environment in which the user can purchase the product naturally without feeling uncomfortable compared to the conventional face-to-face sales at the store. (For example, Patent Document 1). In such a system, a dialogue regarding product purchase is performed with the user by asking the user a question such as "How many pixels is required for the digital camera?".

International Publication No. 2001/084394

According to the conventional technology, a question is asked to the user to confirm the value of the spec of the transaction target category such as a digital camera, and the value of the spec is acquired by the answer of the user to the question.

However, it is hard to say that the conventional technology is highly convenient for users. For example, in the method of directly confirming the spec value with the user, it may not be possible to obtain an appropriate spec value for a user whose use case such as the intended use of the transaction is determined. Therefore, it is desired to appropriately estimate the value of the specifications according to the use case by the user.

Therefore, in this disclosure, we propose an information processing device and an information processing method that can appropriately estimate the value of the specifications according to the use case by the user.

In order to solve the above-mentioned problems, the information processing apparatus of one form according to the present disclosure includes an acquisition unit that acquires character information indicating a use case specified by a user, and the use case based on the character information. Among the specifications of the corresponding transaction target category, an estimation unit for estimating the recommended value of the corresponding specifications corresponding to the use case is provided.

It is a figure which shows an example of information processing which concerns on embodiment of this disclosure. It is a figure which shows an example of the display of the information which concerns on embodiment of this disclosure. It is a figure which shows the structural example of the information processing apparatus which concerns on embodiment of this disclosure. It is a figure which shows an example of the model information storage part which concerns on embodiment of this disclosure. It is a figure which shows an example of the transaction target information storage part which concerns on embodiment of this disclosure. It is a figure which shows an example of the spec list information storage part which concerns on embodiment of this disclosure. It is a figure which shows an example of the model which concerns on embodiment of this disclosure. It is a flowchart which shows the procedure of information processing which concerns on embodiment of this disclosure. It is a figure which shows an example of the flow of recommended processing. It is a figure which shows an example of the flow of the learning process of a model. It is a figure which shows an example of the flow of response processing using a model. It is a figure which shows an example of the learning process using a log. It is a figure which shows an example of the learning process using a log. It is a figure which shows the structural example of the information processing system which concerns on the modification of this disclosure. It is a figure which shows the structural example of the information processing apparatus which concerns on the modification of this disclosure. It is a hardware block diagram which shows an example of the computer which realizes the function of an information processing apparatus.

Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. The information processing apparatus and information processing method according to the present application are not limited by this embodiment. Further, in each of the following embodiments, duplicate description will be omitted by assigning the same reference numerals to the same parts.

The present disclosure will be described according to the order of items shown below.
1. 1. Embodiment 1-1. Outline of information processing according to the embodiment of the present disclosure 1-1-1. Background and effects 1-1-2. Model generation 1-2. Configuration of Information Processing Device According to Embodiment 1-2-1. Model example 1-3. Information processing procedure according to the embodiment 1-4. Processing flow example 1-4-1. Recommended processing flow example 1-4-2. Example of flow of model learning process 1-4-3. Flow example of applied processing for user input 1-5. Learning process example 1-5-1. Learning process example using logs Part 1
1-5-1-1. How to select utterances 1-5-2. Learning process example using logs Part 2
1-5-2-1. How to select a group 2. Other Embodiments 2-1. Modification example 2-2. Other configuration examples 2-3. Others 3. Effect of this disclosure 4. Hardware configuration

[1. Embodiment]
[1-1. Outline of information processing according to the embodiment of the present disclosure]
An outline of information processing will be described with reference to FIGS. 1 and 2. FIG. 1 is a diagram showing an example of information processing according to the embodiment of the present disclosure. FIG. 2 is a diagram showing an example of displaying information according to the embodiment of the present disclosure. Specifically, FIG. 1 is a diagram showing an example of a process (estimation process) of estimating a value of a spec corresponding to a use case by the user by a dialogue with the user. Further, FIG. 1 shows a case where a transaction target such as a product or a service is recommended to the user by using the estimated value. Further, FIG. 2 is a diagram showing an example of displaying information that recommends a transaction target to a user (hereinafter, also referred to as “recommended information”).

Here, specific products and services that are actually traded in commercial transactions are referred to as "transaction targets", and the term "transaction target category" is used as a superordinate concept of the transaction target. The transaction target category is a general term (general name) of the transaction target group belonging to the transaction target category. The transaction target category corresponds to the so-called general names such as the above-mentioned cameras, automobiles, and apples, and the transaction target corresponds to the specific name (product name, service name, etc.) of the transaction target to be actually traded. Each transaction target belongs to one of the transaction target categories. For example, the camera CA1 which is a product (transaction target) actually traded and the product (transaction target) camera CA2 which is actually traded belong to the transaction target category "camera".

In addition, the use case referred to here is a concept relating to a user who uses the transaction target, such as the usage of the transaction target by the user and the user's situation. For example, the use case shows the usage scene of the transaction target and the usage status of the user for the transaction target (for example, whether it is a beginner or not). Thus, a use case is a context regarding the use of a user's trading object.

The information processing according to the embodiment of the present disclosure is realized by the information processing device 100 shown in FIG. The information processing device 100 shown in FIG. 3 is an example of an information processing device that estimates values of specifications and the like. The information processing device 100 is an information processing device that executes information processing according to the embodiment. The information processing device 100 is a terminal device used by a user. FIG. 1 shows an example in which an information processing device 100, which is a terminal device used by a user, estimates values of specifications and the like. For example, the information processing device 100 is used by users such as smartphones, tablet terminals, smart speakers, notebook PCs (Personal Computers), desktop PCs, mobile phones, PDAs (Personal Digital Assistants), and the like. It may be various devices.

The device for estimating the value of the specifications is not limited to the terminal device used by the user, and may be any device. For example, the information processing device that estimates the value of the specifications and the like may be separate from the terminal device used by the user. The system configuration and the like when estimating the spec values and the like on the server side will be described later.

Hereinafter, FIG. 1 will be specifically described. FIG. 1 describes a case where the value of the specifications corresponding to the use case is estimated based on the utterance that specifies the use case of the user U1 through the dialogue with the user U1.

In the example of FIG. 1, the information processing device 100 has functions such as voice signal processing, voice recognition, utterance semantic analysis, and dialogue control. The information processing device 100 may have a voice recognition function. For example, the information processing apparatus 100 may have functions of natural language understanding (NLU: Natural Language Understanding) and automatic speech recognition (ASR: Automatic Speech Recognition). For example, the information processing device 100 may estimate information about a user's intent (intention) or entity (target) from input information uttered by the user.

Further, the information processing device 100 may acquire information used for estimating a spec value or the like by transmitting and receiving information to and from a voice recognition server having a function of natural language understanding and automatic voice recognition. For example, the information processing device 100 may transmit the detected user's utterance information to the voice recognition server and receive the analysis result of the utterance from the voice recognition server.

First, user U1 says, "I want to take a picture of a child's athletic meet, so I want to buy a camera" (step S11). The information processing device 100 detects the utterance of the user U1 saying, "I want to buy a camera because I want to take a picture of a child's athletic meet." The information processing device 100 accepts an utterance by the user U1 as an input. For example, the information processing device 100 converts utterance information such as "I want to buy a camera because I want to take a picture of a child's athletic meet" into text information. As a result, the information processing device 100 acquires the character information (hereinafter referred to as "character information TX1") that "I want to buy a camera because I want to take a picture of a child's athletic meet."

Then, the information processing device 100 determines whether the user's utterance includes a use case (step S12). The information processing device 100 determines whether the character information TX1 includes information indicating a use case. That is, the information processing device 100 determines whether or not the character string "I want to take a picture of a child's athletic meet, so I want to buy a camera" contains a character string indicating a use case. The information processing apparatus 100 determines whether or not the character information TX1 indicates a use case by using the use case determination model M1 (hereinafter, also referred to as “model M1”) used for determining the presence / absence of the use case.

For example, the model M1 is a model that outputs a score indicating whether or not the character information indicates a use case when the character information is input. For example, the model M1 outputs a higher score as the input character information is more likely to indicate a use case. For example, the model M1 outputs a higher score as the input character information is more likely to indicate a use case and is closer to "1", and the input character information is less likely to indicate a use case to be "0". , And outputs a low score. The above is an example, and the model M1 may be a model of various types. Further, details about learning of the model M1 and the like will be described later.

The information processing device 100 uses the model M1 and a predetermined threshold value (for example, "0.7") to determine whether or not the character information indicates a use case. The information processing device 100 compares the score output by the model M1 by inputting the character information with a predetermined threshold value, and determines that the character information indicates a use case, for example, when the score is equal to or higher than the predetermined threshold value. In FIG. 1, the information processing apparatus 100 determines that the character information TX1 indicates a use case because the score output by the model M1 to which the character information TX1 is input is equal to or higher than a predetermined threshold value. That is, the information processing device 100 determines that the character string "I want to take a picture of a child's athletic meet, so I want to buy a camera" contains a character string indicating a use case.

Then, the information processing device 100 estimates various types of information (step S13). The information processing device 100 estimates that the transaction target category is "camera" because the character information TX1 includes a character string indicating that the target that the user wants to purchase is the "camera". Then, the information processing device 100 estimates the specifications (corresponding specifications) corresponding to the use case "shooting of the athletic meet" because the character information TX1 includes a character string for designating the use case "I want to shoot the athletic meet". do. The information processing device 100 estimates that the corresponding spec of the use case "shooting of the athletic meet" is the spec "AF (autofocus)". Further, it is estimated that the information processing device 100 should have the corresponding spec "AF (autofocus)" in the use case "shooting of the athletic meet". That is, the information processing device 100 estimates that the recommended value of the corresponding spec "AF (autofocus)" of the use case "shooting of the athletic meet" is "yes".

As mentioned above, the character information TX1 includes the character string "I want to shoot an athletic meet" indicating a use case, but does not include anything that specifically specifies the value of the specifications to be traded. In this way, the information processing apparatus 100 estimates the value of the spec based on the character information TX1 that does not specifically include the character string indicating the transaction target. The information processing device 100 estimates the specifications for which the value is to be estimated based on the character information TX1 that does not include the character string indicating the specifications. Then, the information processing apparatus 100 estimates the spec value (recommended value) based on the character information TX1 that does not include the spec value. In the example of FIG. 1, the information processing apparatus 100 acquires a spec list (spec list TB1) corresponding to the transaction target category “camera” from the spec list information storage unit 144 (see FIG. 6). Then, as shown in the estimated spec list STB1-1, the information processing apparatus 100 sets "AF" as the corresponding spec among the specs of the transaction target category "camera" and sets the recommended value of the corresponding spec "AF" to "AF". Yes ”is estimated.

The information processing apparatus 100 may estimate the spec value based on the rule base, or may estimate the spec value using the model learned by machine learning. For example, the information processing apparatus 100 may estimate the value of the specifications by using the rule information group stored in the storage unit 14 (see FIG. 3). For example, the information processing apparatus 100 uses the rule information in which the use case "shooting at the athletic meet" is associated with the spec "AF" and the value "yes" in the rule information group, and uses the use case "at the athletic meet". The recommended value of the corresponding spec "AF" corresponding to "shooting" may be estimated to be "yes".

Further, the information processing apparatus 100 may estimate the recommended value of the spec corresponding to the character information TX1 by using the spec estimation model M2 (hereinafter, also referred to as “model M2”) used for estimating the recommended value of the spec.

For example, the model M2 is a model that outputs a plurality of scores corresponding to each of a plurality of specifications. For example, as shown in FIG. 6, the model M2 is a neural network that outputs a score corresponding to each specification in response to input of character information. In this case, the model M2 outputs a score for the spec corresponding to the use case indicated by the input character information among the plurality of specs. Further, the model M2 outputs information (“Not referred” in FIG. 6) indicating that the specs that are not applicable to the use case indicated by the input character information among the plurality of specifications are not applicable. .. The information processing device 100 inputs character information into the model M2, sets the spec for which the score is output as the corresponding spec, and estimates a recommended value based on the output score. The above is an example, and the model M2 may be a model of various types. Further, details about learning of the model M2 and the like will be described later.

For example, the information processing apparatus 100 may estimate the recommended value of the spec on a rule basis before the generation of the model M2, and may estimate the recommended value of the spec using the model M2 after the generation of the model M2. When the number of specs for which a value is set exceeds a predetermined number (referred to as "3" in FIG. 1) in the spec list, the information processing apparatus 100 extracts a transaction target using the value of the spec. , Make recommendations for trading targets. In this case, as shown in the estimated spec list STB1-1, the information processing apparatus 100 has only the spec "AF" value set among the specs of the transaction target category "camera", and the value is set. Since the number "1" is less than the predetermined number "3", we do not recommend the transaction target. The information processing device 100 extracts a transaction target whose spec "AF" corresponds to "Yes" from the transaction target group TG of the transaction target category "camera" as the relevant transaction target, and the number of the extracted corresponding transaction targets is If it is below the threshold, recommendation information may be generated and the transaction target may be recommended.

Then, the information processing device 100 responds to the utterance of the user U1 saying "I want to take a picture of a child's athletic meet, so I want to buy a camera" (step S14). The information processing device 100 outputs a voice response to the utterance of the user U1 based on the result of voice recognition. The information processing device 100 appropriately uses various techniques related to dialogue control to perform a dialogue with the user U1. For example, the information processing apparatus 100 may control the dialogue with the user U1 by using various dialogue models such as the meaning understanding model MUM and the dialogue strategy model DSM in FIG.

By the above-mentioned processing, the information processing apparatus 100 understands the meaning that the user's utterance is "request transmission". Further, the information processing device 100 estimates that the recommended value of the spec "AF" is "Yes" from the use case "Children's athletic meet". Then, the information processing apparatus 100 determines the next action as "explanation of recommended function" or "question about budget".

The information processing device 100 outputs information regarding the recommended specifications of the estimated specifications and a question requesting the user for the values of other specifications by voice. In the example of FIG. 1, the information processing device 100 responds with a response such as "I think the subject is moving, so a camera with autofocus is recommended. How much is the budget?" Specifically, the information processing device 100 responds to the user U1 with the recommended value of the spec "AF" by responding "I think the subject is moving, so a camera with autofocus is recommended." Notify that it is estimated to be "Yes". Further, the information processing apparatus 100 requests the user U1 to specify the value of the spec "price" by responding "How much is the budget?".

Then, the user U1 who recognizes the response from the information processing device 100 says, "I am thinking about 50,000 yen. I also want something that is easy to handle even for beginners" (step S15). The information processing device 100 detects the utterance of the user U1 saying, "I am thinking about 50,000 yen. I also want something that is easy to handle even for beginners." As a result, the information processing device 100 acquires the character information (hereinafter referred to as "character information TX2") that "I am thinking about 50,000 yen. I also want something that is easy to handle even for beginners." In the example of Fig. 1, "I am thinking about 50,000 yen. I also want something that is easy for beginners to handle." Is shown as one character information, but "I am thinking about 50,000 yen." It may be divided into two character information, "Masu." And "I would like something that is easy to handle even for beginners."

Then, the information processing device 100 determines whether the user's utterance includes a use case (step S16). The information processing device 100 determines whether the character information TX2 includes information indicating a use case. That is, the information processing device 100 determines whether or not a character string indicating a use case is included in the character strings "I am thinking about 50,000 yen." And "I would like something that is easy to handle even for beginners." .. The information processing apparatus 100 uses the model M1 to determine whether or not the character information TX2 indicates a use case. In FIG. 1, the information processing apparatus 100 determines that the character information TX2 indicates a use case because the score output by the model M1 to which the character information TX2 is input is equal to or higher than a predetermined threshold value. That is, the information processing device 100 determines that the character string "I am thinking about 50,000 yen. I also want something that is easy for beginners to handle." Contains a character string indicating a use case.

Then, the information processing device 100 estimates various types of information (step S17). Since the character information TX2 includes a character string that specifies a use case of "beginner", the information processing device 100 estimates specifications (corresponding specifications) corresponding to the use case "beginner". The information processing device 100 estimates that the corresponding specifications of the use case "beginner" are the specifications "weight". Further, it is estimated that the information processing apparatus 100 should have a lighter corresponding specification "weight" when used by the use case "beginner". That is, the information processing apparatus 100 estimates that the recommended value of the corresponding specification "weight" of the use case "beginner" is "700 g or less".

In the example of FIG. 1, the information processing apparatus 100 updates the estimated spec list STB1-1 to the estimated spec list STB1-2. Specifically, as shown in the post-estimation spec list STB1-2, the information processing apparatus 100 has the corresponding spec "weight" with "weight" as the corresponding spec among the specs of the transaction target category "camera". The recommended value of is estimated to be "700 g or less".

Further, in the information processing device 100, since the character information TX2 includes a character string for designating the value of the spec "price" of "50,000 yen", the recommended value of the spec "price" is set to "50,000 yen". Set. Specifically, the information processing apparatus 100 sets the recommended value of the spec "price" to "50,000 yen" among the specs of the transaction target category "camera" as shown in the estimated spec list STB1-2. .. The information processing apparatus 100 may estimate the recommended value "700 g or less" of the spec "weight" and the recommended value "50,000 yen" of the spec "price" on a rule basis, or use the model M2. You may estimate.

Then, the information processing apparatus 100 compares the number of specs for which the value is set with the predetermined number in the spec list. In this case, as shown in the estimated spec list STB1-2, the information processing apparatus 100 is set with the values of the specs "AF", "price", and "weight" among the specs of the transaction target category "camera". Since the number "3" of the spec for which the value is set is equal to or more than the predetermined number "3", the transaction target is extracted using the value of the spec (step S18). The information processing device 100 has a spec "AF" of "Yes", a spec "Price" of "50,000 yen", and a spec "Weight" of "700 g or less" from the transaction target group TG of the transaction target category "Camera". Extract the relevant transaction target as the relevant transaction target. In the example of FIG. 1, the information processing apparatus 100 extracts a transaction target including cameras CA1, CA2, CA3, etc. as the corresponding transaction target CTG. If the number of transaction targets in the relevant transaction target CTG is larger than the threshold value, the information processing device 100 may further ask the user a question and further narrow down the corresponding transaction target before performing the processing after step S19 below. good. In the following, the information processing apparatus 100 will be described assuming that the number of transaction targets in the relevant transaction target CTG is equal to or less than the threshold value.

The information processing device 100 classifies the extracted corresponding transaction target CTG (step S19). The information processing device 100 classifies the relevant transaction target CTG based on the similarity of the relevant transaction target CTG or the purchase history of the relevant transaction target CTG by a similar user similar to the user U1.

For example, the information processing device 100 classifies the corresponding transaction target CTG based on the rules. In the example of FIG. 1, the information processing apparatus 100 classifies the corresponding transaction target CTG into three groups A, B, and C, as shown in the classification result CR1. The number of groups to be classified is not limited to three, and may be two or four or more. In addition, each applicable transaction target may belong to two or more groups.

Group A is classified as a transaction target with the spec "camera shake correction". If the recommended value of the spec "camera shake correction" is estimated to be "yes" from the use case "beginner", group A may have a different axis (reference). In group B, transaction targets purchased by users similar to user U1 are classified. In this case, the information processing device 100 classifies the transaction target using the attribute information and purchase history information of each user stored in the storage unit 14. The information processing apparatus 100 classifies the transaction target purchased by a user having similar attributes (age, gender, etc.) of the user U1 among the relevant transaction target CTGs into group B.

Group C is classified into transaction targets that have the abstract characteristic of being easy to carry. For example, each transaction target may be assigned a tag indicating an abstract characteristic such as "easy to carry" in addition to the specifications. In this case, the corresponding transaction target to which the tag "easy to carry" is assigned is classified into group C.

Note that the above is an example, and the axes (references) of each group may be set randomly, or may be set by the administrator of the information processing apparatus 100 or the like. Further, the information processing apparatus 100 is not limited to the above rule-based classification, and various clustering methods may be appropriately used to classify the relevant transaction target CTG. For example, the information processing apparatus 100 may classify the relevant transaction target CTG by using a classification model for classifying the transaction target.

The information processing apparatus 100 generates the classification result CR1 by classifying the camera CA3 and the like into the group A, the camera CA1 and the like into the group B, and the camera CA2 and the like into the group C among the corresponding transaction target CTGs. do. The information processing device 100 generates recommended information to be presented to the user U1 based on the classification result CR1.

Then, the information processing device 100 responds to the utterance of the user U1 saying "I understand. If so, the following cameras are recommended!" (Step S20). The information processing device 100 notifies the user U1 of presenting the recommended transaction target.

By the above-mentioned processing, the information processing apparatus 100 understands the meaning that the user's utterance is "request transmission". Further, the information processing apparatus 100 estimates that the recommended value of the spec "weight" is "700 g or less" from the use case "beginner". Further, the information processing apparatus 100 estimates that the recommended value of the spec "price" is "50,000 yen". Then, the information processing device 100 determines the next action as "end of dialogue, recommendation of recommended product".

Then, the information processing device 100 presents the user with recommended information that recommends the purchase of the transaction target to the user (step S21). The information processing device 100 presents recommended information to the user U1 based on the classification result CR1. As shown in FIG. 2, the information processing apparatus 100 presents the recommended information CT 11 to the user U1 by displaying the recommended information CT 11 on the display unit 16.

FIG. 2 shows an example when the recommended information CT11 is displayed on the display unit 16 which is the display of the information processing device 100. The information processing device 100 is "a camera recommended for U1!" I want to take pictures of children at an athletic meet ", so I recommend a" camera with autofocus and 700g or less "! ", Display information indicating the recommended reason for the transaction target. In addition, the information processing device 100 provides information indicating that the classification standard of Group A is with image stabilization, saying, "Group A: Mr. U1 is a camera beginner, so it is a product group with image stabilization." indicate. Further, the information processing apparatus 100 displays information indicating that the classification standard of group B is the purchase of similar users, such as "Group B: A group of products purchased by a person who is very similar to Mr. U1." In addition, the information processing device 100 displays information indicating that the classification standard of Group C is easy to carry, such as "Group C: A group of products that are easy to carry when going on a trip." In this way, the information processing device 100 presents cameras that meet the conditions of "with autofocus, price of about 50,000 yen, and weight of 700 g or less" in three groups, and also displays explanations of each group. do. As a result, the information processing apparatus 100 can recommend to the user U1 a transaction target suitable for the use case by the user U1.

As described above, the information processing apparatus 100 estimates the recommended value of the specifications according to the use case from the utterance indicating the use case of the user U1. Specifically, the information processing apparatus 100 estimates that the recommended value of the corresponding spec "AF" corresponding to the use case "shooting of the athletic meet" of the user U1 is "yes". In this way, the information processing apparatus 100 can appropriately estimate the value of the specifications according to the use case by the user. In the example of FIG. 1, the case where the character information including the character string "camera" indicating the transaction target category is processed is shown, but the information processing apparatus 100 does not include the transaction target category. May be processed. For example, the information processing apparatus 100 may estimate the transaction target category corresponding to the use case when the character information indicating the use case does not include the character string indicating the transaction target category. For example, the information processing device 100 may presume that the transaction target category corresponding to the use case is "camera" or "video camera" based on the text information "I want to take a picture of a child's athletic meet". good. The information processing apparatus 100 may estimate the transaction target category corresponding to the use case indicated by the character information by using the transaction target rule information in which the use case and the transaction target category are associated with each other. For example, the information processing device 100 estimates the transaction target category using the transaction target rule information stored in the storage unit 14. The information processing apparatus 100 may estimate the transaction target category by using the transaction target rule information in which the use case “shooting of the athletic meet” is associated with the transaction target categories “camera” and “video camera”. In this case, the information processing device 100 estimates that the transaction target category corresponding to the use case indicated by the text information "I want to take a picture of a child's athletic meet" is "camera" or "video camera". The above is an example, and the information processing apparatus 100 may estimate the transaction target category by using a model for estimating the transaction target category (category estimation model). For example, the information processing apparatus 100 uses a category estimation model that outputs information indicating a transaction target category corresponding to the use case of the character information when character information indicating the use case is input, and sets the transaction target category. You may estimate.

[1-1-1. Background and effects, etc.]
For example, it is necessary to specify a transaction target such as a product on a spec basis such as specifying a specific value of a spec, but it is difficult for a user who does not have knowledge of the transaction target to specify a transaction target on the spec base. For example, it is difficult for a user who does not have knowledge of a camera (transaction target category) to specify "AF", "focal length", and the like. In addition, since the user cannot know the reason why the transaction target is recommended by simply presenting the transaction target according to the user's designation, it is not possible to know from the user's point of view whether the transaction target is really suitable for him / her. ..

On the other hand, the information processing device 100 can specify a transaction target on a use case basis, such as specifying a use case. Further, the information processing device 100 has a function of feeding back the reason for recommending the transaction target to the user. Further, in the information processing apparatus 100, it is expected that the product group and the dialogue content recommended by the reinforcement learning mechanism described later will be optimized to be more preferable.

As described above, the information processing apparatus 100 can map use cases to specifications (values). For example, the information processing device 100 maps the use case "I want to take a picture of a child's athletic meet" to the value "Yes" of the spec "AF (autofocus)", and sets the use case "beginner" to the value "weight" of the spec "weight". It can be mapped to "700 g or less (light)". In this way, the information processing apparatus 100 can estimate the value of the specifications from the use case.

In addition, the information processing device 100 presents to the user the reason why the transaction target is recommended (recommended). In the example of FIG. 1, the information processing apparatus 100 groups the relevant transaction targets and recommends the transaction targets (products). In addition, the information processing apparatus 100 also presents the reasons for recommendation.

Although the details will be described later, the information processing device 100 updates the model from the log by reinforcement learning to make it smarter. For example, the information processing device 100 becomes smarter because the presentation of the utterance content and the recommended transaction target (product) group becomes appropriate for the user. The transaction target may include various targets. For example, the transaction target may include human resources (people) in a human resources introduction service or the like.

[1-1-2. Model generation]
Here, the generation of a use case determination model such as model M1 and a spec estimation model such as model M2 will be described. First, the information processing apparatus 100 describes the generation of the use case determination model.

The information processing device 100 uses the information of the utterance case to generate a use case determination model such as the model M1. For example, the information processing device 100 generates the model M1 by using the information of the utterance case stored in the case information storage unit 141. The information processing device 100 uses a utterance case indicating a use case (hereinafter, also referred to as “corresponding utterance case”) as a positive example and a utterance case that does not indicate a use case (hereinafter, also referred to as “non-applicable utterance case”) as a negative example. Generate model M1. For example, the information processing apparatus 100 performs learning processing so that the model M1 outputs a score "1" when the corresponding utterance case (character information) which is a correct example is input to the model M1. To generate. For example, the information processing device 100 performs learning processing so that the model M1 outputs a score "0" when a negative example of non-corresponding utterance case (character information) is input to the model M1. Generate M1. The learning of the use case determination model described above is an example, and the information processing apparatus 100 may learn the use case determination model such as the model M1 by appropriately using various learning methods.

The information processing device 100 uses learning information in which a corresponding utterance case indicating a use case is associated with a spec corresponding to the utterance case, a value of the spec, and the like (hereinafter, also referred to as “correct answer information”). Generate a spec estimation model like M2. For example, the information processing device 100 generates the model M2 by using the learning information stored in the storage unit 14. When the corresponding utterance case (character information) indicating the use case is input to the model M2, the information processing device 100 performs learning processing so that the model M2 outputs the correct answer information corresponding to the corresponding utterance case. , Generate model M2.

The learning of the spec estimation model described above is an example, and the information processing apparatus 100 may learn a spec estimation model such as the model M2 by appropriately using various learning methods. The spec estimation model such as the model M2 may be common to all transaction target categories, or may be generated for each transaction target category. For example, a plurality of transaction target categories may be used for each transaction target category. In this case, a spec estimation model such as model M2 may be generated and used for each transaction target category such as the transaction target category “camera” and “automobile”.

Further, the spec estimation model may be used in common for a plurality of spec common categories when there are a plurality of categories with common specs (also referred to as "spec common category"). The spec common category may be a transaction target category having a predetermined ratio (for example, 50%, 75%, etc.) or more of the common spec. For example, the transaction target category "camera" and the transaction target category "video camera" are at least a predetermined ratio of common specifications, and may be a common specifications category. In this case, the information processing apparatus 100 may generate a spec estimation model that can handle both the transaction target category “camera” and the transaction target category “video camera”. The above is an example, and the spec estimation model may have various forms.

When the information processing apparatus 100 does not generate the use case determination model or the spec estimation model, the information processing apparatus 100 can generate the use case determination model or the spec estimation model from an external model generator that generates the use case determination model or the spec estimation model. You may get it. For example, the information processing apparatus 100 may acquire a use case determination model or a spec estimation model that can correspond to a language (target language) to be estimated such as a spec value from a model generation device. For example, the information processing apparatus 100 may request a use case determination model or a spec estimation model from the model generator, and acquire the use case determination model or the spec estimation model of the target language from the model generator.

[1-2. Configuration of Information Processing Device According to Embodiment]
Next, the configuration of the information processing device 100, which is an example of the information processing device that executes the information processing according to the embodiment, will be described. FIG. 3 is a diagram showing a configuration example of the information processing device 100 according to the embodiment of the present disclosure. For example, the information processing device 100 shown in FIG. 3 is an example of the information processing device. The information processing device 100 is a computer that realizes a function as an information processing device described later.

As shown in FIG. 3, the information processing device 100 includes a communication unit 11, an input unit 12, an audio output unit 13, a storage unit 14, a control unit 15, and a display unit 16. In the example of FIG. 3, the information processing device 100 includes an input unit 12 (for example, a keyboard, a mouse, etc.) that receives various operations from the administrator of the information processing device 100, and a display unit 16 (for example, a display unit 16) for displaying various information. , Liquid crystal display, etc.).

The communication unit 11 is realized by, for example, a NIC (Network Interface Card), a communication circuit, or the like. The communication unit 11 is connected to the communication network N (network such as the Internet) by wire or wirelessly, and transmits / receives information to / from other devices via the communication network N.

The input unit 12 accepts various inputs. The input unit 12 receives the detection by the sensor as an input. The input unit 12 receives sound as input by a sound sensor having a function of detecting voice. The input unit 12 receives the voice information detected by the microphone that detects the voice as the input information. The input unit 12 receives the voice spoken by the user as input information. Various operations are input from the user to the input unit 12. The input unit 12 accepts input by the user. The input unit 12 may accept the user's selection of the learning method. The input unit 12 may accept various operations from the user via a keyboard, mouse, or touch panel provided in the information processing device 100.

The voice output unit 13 outputs information by voice. The audio output unit 13 is realized by a speaker that outputs audio. The voice output unit 13 outputs voice according to the control of the dialogue management unit 153. The voice output unit 13 outputs a response corresponding to the user's utterance by voice under the control of the dialogue management unit 153.

The storage unit 14 is realized by, for example, a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory (Flash Memory), or a storage device such as a hard disk or an optical disk. The storage unit 14 includes a case information storage unit 141, a model information storage unit 142, a transaction target information storage unit 143, and a spec list information storage unit 144.

The storage unit 14 stores not only the information stored in the case information storage unit 141, the model information storage unit 142, the transaction target information storage unit 143, and the spec list information storage unit 144, but also various types of information. .. For example, the storage unit 14 stores information of a voice recognition application (program) that realizes a voice recognition function. For example, the terminal device 10 performs voice recognition processing by a voice recognition application. The storage unit 14 stores various information used for displaying the information. The storage unit 14 stores various information used for voice recognition.

Although not shown, the case information storage unit 141 stores information related to the utterance case. For example, the case information storage unit 141 can distinguish between an utterance case indicating a use case (corresponding utterance case) and an utterance case not indicating a use case (non-applicable utterance case). For example, the case information storage unit 141 memorizes the corresponding utterance case indicating the use case and the non-applicable utterance case not indicating the use case by associating them with different flags. The case information storage unit 141 stores the corresponding utterance case as a positive example of the learning data of the use case determination model, and stores the non-corresponding utterance case as a negative example of the learning data of the use case determination model.

For example, the case information storage unit 141 stores learning information used for learning the spec estimation model. The case information storage unit 141 stores the value of the spec estimated from the utterance, the spec, and the transaction target category in association with the corresponding utterance case indicating the use case. The case information storage unit 141 may separately store the corresponding utterance case showing the use case and the non-corresponding utterance case not showing the use case. For example, the case information storage unit 141 stores the information of the use case & slot correspondence database USD described later as the information of the corresponding utterance case. Further, the case information storage unit 141 stores the information of the Not use case utterance case NUD, which will be described later, as the information of the non-applicable utterance case.

The model information storage unit 142 according to the embodiment stores information about the model. For example, the model information storage unit 142 stores information (model data) indicating the structure of the model (network). FIG. 4 is a diagram showing an example of a model information storage unit according to the embodiment of the present disclosure. FIG. 4 shows an example of the model information storage unit 142 according to the embodiment. In the example shown in FIG. 4, the model information storage unit 142 includes items such as "model ID", "use", and "model data".

"Model ID" indicates identification information for identifying the model. "Use" indicates the use of the corresponding model. "Model data" indicates model data. In FIG. 4, an example in which conceptual information such as "MDT1" is stored in "model data" is shown, but in reality, various information constituting the model such as information and functions related to the network included in the model are stored. included.

In the example shown in FIG. 4, the model (model M1) identified by the model ID "M1" indicates that the use is "use case determination". Model M1 indicates that it is a use case determination model used for estimating the value of the spec. Further, it is shown that the model data of the model M1 is the model data MDT1.

In the example shown in FIG. 4, the model (model M2) identified by the model ID "M2" indicates that the use is "spec estimation". Model M2 indicates that it is a spec estimation model used for estimating spec values. Further, it is shown that the model data of the model M2 is the model data MDT2. Although only the models M1 and M2 are shown in FIG. 4, the model information storage unit 142 may store models other than the models M1 and M2.

The model information storage unit 142 is not limited to the above, and may store various information depending on the purpose. For example, the model information storage unit 142 stores the model information learned (generated) by the learning process. The model information storage unit 142 stores the parameter information of the models M1 and M2 learned (generated) by the learning process.

The transaction target information storage unit 143 stores various information related to the transaction target. FIG. 5 is a diagram showing an example of a transaction target information storage unit according to the embodiment of the present disclosure. For example, the transaction target information storage unit 143 stores various information related to various transaction targets such as goods and services. An example of the transaction target information storage unit 143 according to the embodiment is shown. In the example of FIG. 5, the case information storage unit 141 includes items such as "transaction target ID", "transaction target", "transaction target category", and "spec information".

"Transaction target ID" indicates identification information for identifying a transaction target. Further, "transaction target" indicates a transaction target corresponding to the transaction target ID. In the example of FIG. 5, the transaction target is indicated by an abstract code such as “camera CA1”, but the transaction target is a specific product (commodity) or service.

In addition, the "transaction target category" indicates the transaction target category to which the corresponding transaction target belongs. "Spec information" indicates information on the corresponding specifications of the transaction target. In the example of FIG. 5, the spec information is indicated by an abstract code such as "SINF1", but the spec information includes various information indicating specifications such as the price and properties (performance and function) of the transaction target. For example, the spec information is specific information of various specifications such as the number of pixels, the weight, and the price of the transaction target when the transaction target category is "camera".

In the example of FIG. 5, it is shown that the transaction target (transaction target TT1) identified by the transaction target ID “TT1” is the camera CA1. Camera CA1 indicates that it belongs to the transaction target category "camera". The spec information of the camera CA1 indicates that it is "spec information SINF1". Spec information SINF1 includes specific information of various specifications such as the number of pixels of the camera CA1 "10 million", the weight "500 g", and the price "100,000 yen".

The transaction target information storage unit 143 is not limited to the above, and may store various information depending on the purpose. The transaction target information storage unit 143 may store the transaction target by dividing the spec list (table) for each transaction target category.

The spec list information storage unit 144 according to the embodiment stores various information related to the spec list. The spec list information storage unit 144 stores various information related to the spec list for each transaction target category. FIG. 6 is a diagram showing an example of a spec list information storage unit according to the embodiment of the present disclosure.

In the example of FIG. 6, the spec list information storage unit 144 stores information (spec list) for each transaction target category such as spec list information TB1 and spec list information TB2. For example, the spec list information TB1 indicates information related to the spec list of the transaction target category “camera”. Further, for example, the spec list information TB2 indicates information related to the spec list of the transaction target category “automobile”.

The spec list information TB1 and the spec list information TB2 shown in FIG. 6 include items such as "transaction target category" and "spec". Further, the case where the "spec" includes items such as "# 1", "# 2", "# 3", and "# 4" is illustrated. The "specs" are not limited to "# 1", "# 2", "# 3", and "# 4", but include "# 5", "# 6", and other numbers corresponding to the specifications. May be included. In addition, the parentheses in the items of each spec indicate the specific spec (name).

In the example of FIG. 6, the spec list information TB1 has the specs “# 1 (AF)”, “# 2 (pixel count)”, “# 3 (price)”, and “# 4 (# 4 (price)” for the transaction target category “camera”. Weight) ”is included. That is, it is shown that the specifications of the transaction target category "camera" include specifications such as AF (autofocus), the number of pixels, price, and weight.

The spec list information storage unit 144 is not limited to the above, and may store various information depending on the purpose. For example, the spec list information storage unit 144 may store information regarding one spec list, which is a union of specs of all transaction target categories.

Return to Fig. 3 and continue the explanation. In the control unit 15, for example, a program (for example, an information processing program according to the present disclosure) stored inside the information processing apparatus 100 by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like is stored in a RAM (Random Access Memory). ) Etc. are executed as a work area. Further, the control unit 15 is a controller, and may be realized by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

As shown in FIG. 3, the control unit 15 includes an acquisition unit 151, a learning unit 152, a dialogue management unit 153, an estimation unit 154, an extraction unit 155, a generation unit 156, and a transmission unit 157. , Realize or execute the information processing functions and actions described below. The internal configuration of the control unit 15 is not limited to the configuration shown in FIG. 3, and may be another configuration as long as it is a configuration for performing information processing described later.

The acquisition unit 151 acquires various information. The acquisition unit 151 acquires various information from an external information processing device. The acquisition unit 151 acquires various information from the storage unit 14. The acquisition unit 151 acquires the information received by the input unit 12.

The acquisition unit 151 acquires various information from the storage unit 14. The acquisition unit 151 acquires various information from the case information storage unit 141, the model information storage unit 142, the transaction target information storage unit 143, and the spec list information storage unit 144. The acquisition unit 151 acquires learning data. The acquisition unit 151 acquires dictionary information from the case information storage unit 141.

The acquisition unit 151 may acquire the model. The acquisition unit 151 may acquire information indicating the network structure of the model. The acquisition unit 151 acquires a model from an external information processing device or a storage unit 14 that provides the model. For example, the acquisition unit 151 acquires the model M1 and the model M2 from the model information storage unit 142. For example, the acquisition unit 151 acquires information indicating the network structure of the model M1 and the model M2 from the model information storage unit 142. The acquisition unit 151 acquires learning data used for learning a model by machine learning. The acquisition unit 151 acquires learning data used for learning the model from the case information storage unit 141.

The acquisition unit 151 acquires various information learned by the learning unit 152. The acquisition unit 151 acquires various information estimated by the estimation unit 154. The acquisition unit 151 acquires the information extracted by the extraction unit 155. The acquisition unit 151 acquires various information generated by the generation unit 156.

The acquisition unit 151 acquires character information indicating a use case specified by the user. The acquisition unit 151 acquires character information based on the user's utterance. The acquisition unit 151 acquires character information that does not include the value of the corresponding specification. The acquisition unit 151 acquires character information that does not include a character string indicating the specifications. The acquisition unit 151 acquires character information that does not include a character string indicating a transaction target category. The acquisition unit 151 acquires character information that does not include a character string indicating a transaction target belonging to the transaction target category. The acquisition unit 151 acquires character information indicating the use of the transaction target category. The acquisition unit 151 acquires character information indicating a usage scene of the transaction target category. The acquisition unit 151 acquires character information indicating the user's situation. The acquisition unit 151 acquires character information indicating the usage status of the user with respect to the transaction target of the transaction target category. The acquisition unit 151 acquires character information indicating whether or not the user is a beginner regarding the use of the transaction target of the transaction target category.

The learning unit 152 performs the learning process. The learning unit 152 performs various learning. The learning unit 152 learns various types of information based on the information acquired by the acquisition unit 151. The learning unit 152 learns (generates) a model. The learning unit 152 learns various information such as a model. The learning unit 152 generates a model by learning. The learning unit 152 learns a model by using various techniques related to machine learning. For example, the learning unit 152 learns the parameters of the model (network). The learning unit 152 learns a model by using various techniques related to machine learning.

The learning unit 152 performs various learning. The learning unit 152 learns various types of information based on the information stored in the storage unit 14. The learning unit 152 learns the model based on the information stored in the case information storage unit 141 and the model information storage unit 142.

Learning unit 152 learns network parameters. For example, the learning unit 152 learns the network parameters of the model M1 and the model M2. The learning unit 152 learns the model M1 and the model M2 by learning the network parameters of the model M1 and the model M2.

The learning unit 152 learns the model using the learning data which is a combination of the learning byte string and the correct answer information corresponding to the learning byte string. The learning unit 152 learns the model corresponding to the language by using the learning data of the language corresponding to the character string. The learning unit 152 generates a model by performing learning processing based on the learning data (teacher data) stored in the case information storage unit 141. The learning unit 152 generates a model by performing a learning process using the learning data stored in the case information storage unit 141. For example, the learning unit 152 generates a model used for estimating a spec value or the like. The learning unit 152 learns the network parameters of the model M1 and the model M2, and generates the model M1 and the model M2.

The learning method by the learning unit 152 is not particularly limited. For example, learning data in which a byte string corresponding to a character string and a probability distribution of the character string are linked is prepared, and the learning data is used as a multi-layer neural network. You may learn by inputting into the calculation model based on. Further, for example, a method based on DNN (Deep Neural Network) such as CNN (Convolutional Neural Network) or 3D-CNN may be used. The learning unit 152 may use a method based on a recurrent neural network (RNN) or an LSTM (Long Short-Term Memory units) that extends the RNN.

The learning unit 152 stores the model generated by learning in the model information storage unit 142. The learning unit 152 generates the model M1 and the model M2. In this case, the learning unit 152 stores the generated model M1 and model M2 in the model information storage unit 142. The learning unit 152 learns a model based on each data used as learning data and correct answer information.

Dialogue management unit 153 manages voice dialogue. The dialogue management unit 153 executes a process related to the voice dialogue. The dialogue management unit 153 executes voice dialogue control. The dialogue management unit 153 realizes a voice recognition function. The dialogue management unit 153 executes speech dialogue control by appropriately using techniques such as natural language understanding (NLU) and automatic speech recognition (ASR). The dialogue management unit 153 executes a voice dialogue with the user. The dialogue management unit 153 controls the voice output unit 13 to output voice. The dialogue management unit 153 causes the voice output unit 13 to output a response corresponding to the user's utterance by voice. The dialogue management unit 153 may be integrated with the estimation unit 154.

The estimation unit 154 performs estimation processing. The estimation unit 154 estimates various types of information. The estimation unit 154 estimates various types of information based on the information acquired from the external information processing device. The estimation unit 154 estimates various types of information based on the information stored in the storage unit 14. The estimation unit 154 estimates various types of information based on the information stored in the case information storage unit 141, the model information storage unit 142, the transaction target information storage unit 143, and the spec list information storage unit 144.

The estimation unit 154 performs analysis processing. The estimation unit 154 analyzes various information. The estimation unit 154 analyzes various information based on the information acquired from the external information processing device. The estimation unit 154 analyzes various information based on the information stored in the storage unit 14. The estimation unit 154 analyzes various information based on the information stored in the case information storage unit 141, the model information storage unit 142, the transaction target information storage unit 143, and the spec list information storage unit 144.

The estimation unit 154 analyzes the character information corresponding to the user's utterance by appropriately using a natural language processing technique such as morphological analysis. The estimation unit 154 estimates (identifies) the content of the user's utterance by semantic analysis using the character information corresponding to the user's utterance. The estimation unit 154 estimates (identifies) the content of the character information by analyzing the character information by appropriately using semantic analysis and dialogue state estimation. For example, the estimation unit 154 estimates the content of the user's utterance corresponding to the character information by appropriately analyzing the character information using various conventional techniques such as parsing.

Further, the estimation unit 154 may estimate the content such as the intention of the user's utterance by analyzing the user's utterance. For example, the estimation unit 154 estimates the content such as the intention of the user's utterance by appropriately using various conventional techniques. For example, the estimation unit 154 estimates the content of the user's utterance by analyzing the user's utterance by appropriately using various conventional techniques. For example, the estimation unit 154 extracts important keywords from the character information of the user's utterance, and estimates the content of the user's utterance based on the extracted keywords.

The estimation unit 154 makes various determinations. The estimation unit 154 determines whether or not the character information indicates a use case. The estimation unit 154 makes various determinations based on the information acquired by the acquisition unit 151. The estimation unit 154 determines whether or not the character information indicates a use case based on the model (use case determination model) learned by the learning unit 152.

The estimation unit 154 estimates the recommended value of the corresponding spec corresponding to the use case among the specifications of the transaction target category corresponding to the use case based on the character information. The estimation unit 154 estimates the recommended value by using a model that outputs a score indicating the recommended value in response to the input of character information. The estimation unit 154 uses a model that outputs a plurality of scores corresponding to each of the plurality of specifications, and estimates the recommended value of the specifications for which the model outputs the scores.

The estimation unit 154 estimates the recommended value based on the character information that does not include the value of the corresponding specification. The estimation unit 154 estimates the corresponding specifications for which the recommended value is estimated, based on the character information that does not include the character string indicating the specifications. The estimation unit 154 estimates the transaction target category for which the recommended value is estimated, based on the character information that does not include the character string indicating the transaction target category. The estimation unit 154 estimates the transaction target category for which the recommended value is to be estimated, based on the character string character information indicating the transaction target belonging to the transaction target category.

The estimation unit 154 estimates the recommended value of the corresponding specifications of the transaction target category used for the purpose. The estimation unit 154 estimates the recommended value of the corresponding specifications of the transaction target category suitable for use in the usage scene. The estimation unit 154 estimates the recommended value of the corresponding specifications of the transaction target category corresponding to the user's situation. The estimation unit 154 estimates the recommended value of the corresponding specifications of the transaction target category corresponding to the usage status of the user. When the user is a beginner, the estimation unit 154 estimates the recommended value of the corresponding specifications of the transaction target category to the value corresponding to the beginner. The estimation unit 154 estimates a plurality of recommended values corresponding to each of the plurality of specifications.

Extraction unit 155 performs various extractions. The extraction unit 155 extracts various information based on the information acquired by the acquisition unit 151. The extraction unit 155 extracts various information based on the information estimated by the estimation unit 154. The acquisition unit 151 extracts various information based on the information extracted by the extraction unit 155. The extraction unit 155 extracts various information based on the information stored in the storage unit 14. The extraction unit 155 extracts various information based on the information stored in the case information storage unit 141, the model information storage unit 142, the transaction target information storage unit 143, and the spec list information storage unit 144. The extraction unit 155 extracts various information from the information stored in the storage unit 14. The extraction unit 155 extracts various information from the information stored in the transaction target information storage unit 143.

The extraction unit 155 extracts the transaction target whose corresponding spec value corresponds to the recommended value from the transaction target group of the transaction target category as the relevant transaction target. The extraction unit 155 extracts the corresponding transaction target in which each of the plurality of values of the plurality of specifications corresponds to the plurality of recommended values.

Generation unit 156 performs various generations. The generation unit 156 generates various information based on the information acquired by the acquisition unit 151. The generation unit 156 generates various information based on the information estimated by the estimation unit 154. The acquisition unit 151 generates various information based on the information extracted by the extraction unit 155. The generation unit 156 generates various information based on the information stored in the storage unit 14. The generation unit 156 generates various information based on the information stored in the case information storage unit 141, the model information storage unit 142, the transaction target information storage unit 143, and the spec list information storage unit 144.

The generation unit 156 generates various information to be displayed on the display unit 16. The generation unit 156 may generate various information such as character information to be displayed on the display unit 16 and image information such as a graph. In this case, the generation unit 156 generates information (image) about the screen by appropriately using various conventional techniques related to the image. The generation unit 156 generates an image by appropriately using various conventional techniques related to GUI. For example, the generation unit 156 displays an image in CSS (Cascading Style Sheets), Javascript (registered trademark), HTML (HyperText Markup Language), or any language capable of describing information processing such as the above-mentioned information display and operation reception. It may be generated.

The generation unit 156 generates recommended information that recommends the user to purchase the transaction target based on the corresponding transaction target extracted by the extraction unit 155. The generation unit 156 generates recommended information when the number of applicable transaction targets is equal to or less than the threshold value. The generation unit 156 generates recommended information indicating the classification result of classifying the relevant transaction target. The generation unit 156 classifies the relevant transaction target based on the similarity of the relevant transaction target or the purchase history of the relevant transaction target by a similar user similar to the user, and generates recommended information indicating the classification result.

The transmission unit 157 transmits various information. The transmission unit 157 provides various types of information. The transmission unit 157 provides various information to an external information processing device. The transmission unit 157 transmits various information to an external information processing device. The transmission unit 157 transmits the information stored in the storage unit 14. The transmission unit 157 transmits the information stored in the case information storage unit 141, the model information storage unit 142, the transaction target information storage unit 143, and the spec list information storage unit 144. The transmission unit 157 transmits information on the model learned by the learning unit 152. The transmission unit 157 transmits the estimation result by the estimation unit 154. The acquisition unit 151 transmits the information extracted by the extraction unit 155. The transmission unit 157 transmits the information generated by the generation unit 156.

The display unit 16 displays various information. The display unit 16 is a display device (display unit) such as a display, and displays various information. The display unit 16 displays the information of the estimation result by the estimation unit 154. The display unit 16 displays the information generated by the generation unit 156.

The display unit 16 displays the recommended information generated by the generation unit 156. The display unit 16 displays recommended information indicating the classification result of classifying the relevant transaction target. The display unit 16 classifies the relevant transaction target based on the similarity of the relevant transaction target or the purchase history of the relevant transaction target by a similar user similar to the user, and displays recommended information indicating the classification result. The information processing device 100 does not have to have the display unit 16 when the information processing device 100 outputs only by voice.

[1-2-1. Model example]
Here, an example of the spec estimation model used by the information processing apparatus 100 will be described with reference to FIG. 7. FIG. 7 is a diagram showing an example of a model according to the embodiment of the present disclosure. The model M2, which is a neural network shown in FIG. 7, takes character information as input and outputs a score of specifications corresponding to the character information. In FIG. 7, the input is character information, and the output is a plurality of outputs corresponding to each of a plurality of specifications such as specifications SP1 to SP7. The output is not limited to 7 of the specifications SP1 to SP7, and may be 8 or more, or 6 or less.

The output of spec SP1 corresponds to spec "AF". The output of Spec SP2 corresponds to Spec "SELPHY". The output of Spec SP3 corresponds to the Spec "Time Lapse". The output of Spec SP4 corresponds to the Spec "Lower Budget". The output of Spec SP5 corresponds to the Spec "Budget Upper Limit". The output of the spec SP6 corresponds to the spec "lower limit of weight". The output of the spec SP7 corresponds to the spec "upper limit of weight".

In the example of FIG. 7, the output information OUT1 when the character information IN1 "I want to take a picture of a child at an athletic meet" is input to the model M2 is shown as an example. In this case, the model M2 outputs a value (for example, "1" or the like) corresponding to "Yes" as the output of the spec SP1 as shown in the score SC1. The model M2 outputs a value corresponding to "Not referred" indicating non-target as the output of the specifications SP2 to SP5 as shown in the scores SC2 to SC5. That is, in the example of FIG. 7, in the use case indicated by the character information IN1 "I want to take a picture of a child at an athletic meet", the specs SP2 to SP5 are not subject to estimation. Further, the model M2 outputs a value corresponding to "0 g" as an output of the spec SP6 as shown in the score SC6. The model M2 outputs a value corresponding to "700 g" as an output of the spec SP7 as shown in the score SC7.

As described above, in the example of FIG. 7, the value of the spec "AF" is "Yes" and the value of the spec "Weight" is "0 g to 700 g" from the character information IN1 "I want to take a picture of a child at an athletic meet." It is estimated to be. The above is an example, and the model M2 may have various forms.

[1-3. Information processing procedure according to the embodiment]
Next, the procedure of information processing according to the embodiment will be described with reference to FIG. FIG. 8 is a flowchart showing an information processing procedure according to the embodiment of the present disclosure.

As shown in FIG. 8, the information processing apparatus 100 acquires character information indicating a use case specified by the user (step S101). For example, the information processing device 100 acquires character information based on a user's utterance in response to a dialogue with the user. The information processing apparatus 100 estimates the recommended value of the corresponding spec corresponding to the use case among the specifications of the transaction target category corresponding to the use case based on the character information (step S102). For example, the information processing apparatus 100 estimates the recommended value of the corresponding spec by using the spec list of the transaction target category corresponding to the use case.

[1-4. Processing flow example]
From here, the processing flow will be described with reference to FIGS. 9 to 11.

[1-4-1. Recommended processing flow example]
First, the flow of recommended processing will be described with reference to FIG. FIG. 9 is a diagram showing an example of the flow of recommended processing. Specifically, FIG. 9 is a diagram showing an example of a recommended processing flow of a product according to a dialogue with a user. The same points as those in FIGS. 1 and 2 will be omitted as appropriate.

In FIG. 9, the information processing device 100 accepts user input (step S201). For example, the information processing device 100 acquires the character information input by the user's utterance.

Then, the information processing device 100 executes the dialogue management process (step S202). For example, the information processing apparatus 100 executes the dialogue management process using information such as the meaning understanding model MUM, the use case & slot correspondence database USD, and the dialogue strategy model DSM. For example, the information processing device 100 determines a response corresponding to a user's utterance.

For example, the dialogue management process includes the following three processes from the first process to the third process.

The dialogue management process includes understanding the meaning of the user's utterance as the first process. For example, the first process includes understanding the intention of the user's utterance whether it is a greeting, a question, a request transmission, or the like. For example, the information processing device 100 estimates the specifications when the meaning of the user's utterance conveys the request. Further, the information processing apparatus 100 estimates that the recommended value of the spec "price" is "50,000 yen or less" when the user's utterance is "50,000 yen or less is good." Further, when the user's utterance is "I want to take a picture of a child at an athletic meet", the information processing device 100 estimates that the recommended value of the spec "AF (autofocus)" is "Yes".

The dialogue management process includes updating the internal state as the second process. For example, in the second process, the internal state (spec) is updated with reference to the past history. For example, the information processing apparatus 100 updates the information in the spec list with the value of the spec "AF" to "Yes" and the value of the spec "Price" to "50,000 yen or less".

The dialogue management process includes the determination of actions such as the content of the next utterance and the end of the dialogue as the third process. For example, in the third process, the utterance content is determined based on the internal state. For example, the information processing device 100 determines to give a response confirming the use case to the user, such as "Are you planning to use it for travel or the like?" Further, when the predetermined condition is satisfied, the information processing apparatus 100 ends the dialogue, and shifts to the presentation of the transaction target group such as the recommended product and the explanation of the reason for recommendation.

The presentation of the transaction target group and the explanation of the reason for recommendation include, for example, the following processing. The information processing apparatus 100 clusters and presents about 3 groups based on the narrowed down conditions. Moreover, the method of forming a group may be various aspects. For example, the information processing apparatus 100 classifies based on other specifications such that no value is set. For example, the information processing device 100 is classified into a product group that can shoot a moving image, a product group that is easy to carry, and the like. For example, the information processing apparatus 100 classifies products (transaction targets) purchased by people (similar users) having similar gender, age, personality, etc. into one group. For example, the information processing device 100 classifies by utilizing the information obtained in the dialogue. For example, the information processing device 100 is classified into a product group that is easy to handle even for "beginners".

Then, the information processing device 100 executes the response (step S203). For example, the information processing device 100 outputs a response to a user's utterance as voice.

Then, the information processing device 100 determines whether to end the dialogue (step S204). For example, the information processing apparatus 100 determines whether to end the dialogue by using the number of specs for which the value is set and the number of the corresponding transaction targets extracted based on the spec for which the value is set. For example, the information processing apparatus 100 interacts when the number of specs for which values are set is equal to or more than a predetermined number, or when the number of corresponding transaction targets extracted based on the specs for which values are set is equal to or less than a predetermined number. Is determined to end.

When the information processing device 100 determines that the dialogue is terminated (step S204: Yes), the information processing device 100 recommends the transaction target and presents the explanation of the reason for the recommendation to the user (step S205). On the other hand, when the information processing apparatus 100 determines that the dialogue is not terminated (step S204: No), the information processing apparatus 100 returns to step S201 and repeats the process.

[1-4-2. Model learning process flow example]
Next, the flow of the learning process of the model will be described with reference to FIG. FIG. 10 is a diagram showing an example of the flow of the learning process of the model. For example, FIG. 10 shows an example of data collection and machine learning model creation for responding to use case-based utterances.

As shown in FIG. 10, the information processing apparatus 100 uses and slots by processing the information stored in the log database WD including various Web data such as the knowledge search service and the log information of the purchase consultation thread into text. Generate information on the corresponding database USD (step S301). The log database WD may include various information as well as Web data. The log database WD may include logs related to the user's purchase at the physical store. For example, the log database WD may include actual store data (image, sound, etc.) such as sensor data collected (detected) by a camera, a microphone, or the like worn by a clerk of the actual store. The log database WD may include actual store data in which the sensor data at the time of customer service of the clerk of the actual store is associated with the purchase presence / absence of the user to be served and the transaction target such as the purchased product.

Further, the information processing device 100 may generate information on the Not use case utterance case NUD using the log database WD, or information on the Not use case utterance case NUD from an external device that provides information on past utterance cases. May be obtained.

Then, the information processing apparatus 100 generates a model M1 which is a use case determination model (model # 1) by machine learning (step S302). The information processing apparatus 100 generates the model M1 by using the information of the use case & slot correspondence database USD and the information of the Not use case utterance case NUD. The information processing apparatus 100 generates the model M1 by using the information of the use case & slot correspondence database USD as a positive example and the information of the Not use case utterance case NUD as a negative example. Since this point has been described above, the description thereof will be omitted.

Further, the information processing apparatus 100 generates a model M2 which is a spec estimation model (model # 2) by machine learning (step S303). The information processing apparatus 100 generates the model M2 by using the information of the use case & slot correspondence database USD. The information processing device 100 generates the model M2 using the information of the use case & slot correspondence database USD as learning information. Since this point has been described above, the description thereof will be omitted.

[1-4-3. Flow example of applied processing for user input]
Next, the flow of application processing for user input will be described with reference to FIG. FIG. 11 is a diagram showing an example of the flow of response processing using the model. For example, FIG. 11 shows an example of a processing flow during a dialogue with a user.

In FIG. 11, the information processing device 100 accepts user input (step S401). For example, the information processing device 100 acquires the character information input by the user's utterance.

First, the information processing device 100 determines whether or not the user input indicates a use case (step S402). The information processing device 100 determines whether or not the user's input indicates a use case by using the model M1 which is a use case determination model.

When the information processing apparatus 100 determines that the user input indicates a use case (step S402: Yes), the information processing device 100 estimates the optimum specifications for the use case (step S403). The information processing apparatus 100 estimates the recommended value of the spec corresponding to the use case by using the model M2 which is a spec estimation model.

Then, the information processing device 100 executes the dialogue management process (step S404). For example, the information processing device 100 determines a response corresponding to a user's utterance.

Then, the information processing device 100 executes the response (step S405). For example, the information processing device 100 outputs a response to a user's utterance as voice.

On the other hand, when the information processing apparatus 100 determines that the user input does not indicate a use case (step S402: No), the information processing apparatus 100 executes the dialogue management process (step S406). That is, when the information processing apparatus 100 determines that the user input does not indicate a use case, the information processing apparatus 100 executes the dialogue management process without estimating the value of the specifications. For example, the information processing device 100 determines a response corresponding to a user's utterance.

Then, the information processing device 100 executes the response (step S407). For example, the information processing device 100 outputs a response to a user's utterance as voice.

[1-5. Learning process example]
Next, an example of learning processing will be described with reference to FIGS. 12 and 13. 12 and 13 are diagrams showing an example of learning processing using logs. Specifically, FIG. 12 is a diagram showing an example of a process of reinforcement learning using a log related to utterance selection. Further, FIG. 13 is a diagram showing an example of a process of reinforcement learning using a log related to a transaction target recommendation.

[1-5-1. Learning process example using logs Part 1]
First, FIG. 12 will be described. For example, FIG. 12 shows a mechanism of reinforcement learning using a log for the time of utterance selection.

FIG. 12 shows a case where the spec "AF" is provided, the budget is about 50,000 yen, and the state information INF1 that the utterance intention conveys the request is given. That is, the state information INF1 indicates that the recommended value of the spec "AF" is "Yes" and the recommended value of the spec "Price" is set to "about 50,000 yen". The state information INF1 indicates the specifications currently filled and the intention of the user's utterance immediately before.

Then, the utterance candidate group SG1 and the dialogue end DE1 indicate an option of which utterance to make next or to terminate the dialogue and shift to the recommendation of the transaction target. The utterance candidate group SG1 and the dialogue end DE1 indicate possible action options next. The utterance candidate group SG1 includes a plurality of utterance candidates such as the candidate CD1 which is the utterance candidate # 1, the candidate CD2 which is the utterance candidate # 2, and the candidate CDN which is the utterance candidate #N.

Then, in FIG. 12, the reward is whether or not the user has purchased the recommended transaction target. That is, the reward is given depending on whether or not the finally recommended transaction target has been purchased. In this way, the information processing apparatus 100 performs reinforcement learning based on the state shown in the state information INF1, the action shown in the utterance candidate group SG1 and the dialogue end DE1, and the above reward. For example, the information processing apparatus 100 may determine the next action to be performed by using an action determination model for determining the next action from a plurality of utterance candidates and options such as the end of dialogue. In this case, the information processing device 100 may update the behavior determination model by the reinforcement learning described above. The information processing device 100 is not limited to model-based reinforcement learning, and may perform model-free reinforcement learning and the like.

[1-5-1-1. How to select utterances]
The method of selecting the utterance may be appropriately changed depending on the presence or absence of the log. For example, when the information processing apparatus 100 does not have a log of a predetermined amount or more, the utterance may be determined on a rule basis, or which specifications should be filled may be determined by searching. For example, when there is not enough logs, the information processing device 100 determines an utterance using predetermined rule information, and determines specifications and values to be estimated.

Further, for example, when logs are collected, the information processing device 100 selects an utterance having a high expected value of reward. For example, when the information processing apparatus 100 has a predetermined amount or more of logs, the information processing apparatus 100 selects an utterance having a high expected value of reward. The information processing device 100 may use the above-mentioned action determination model to determine the next action to be performed, such as selection of an utterance or termination of a dialogue.

[1-5-2. Learning process example using logs Part 2]
Next, FIG. 13 will be described. The same points as in FIG. 12 will not be described. For example, FIG. 13 shows a mechanism of reinforcement learning using a log for the time of recommendation of a transaction target.

FIG. 13 shows a case where the spec "AF" is provided, the budget is about 50,000 yen, and the state information INF2 that the acquired information is a beginner is given. That is, the state information INF2 indicates that the recommended value of the spec "AF" is "Yes" and the recommended value of the spec "Price" is set to "about 50,000 yen". Further, the state information INF2 indicates that the user having a dialogue is a beginner. The state information INF2 indicates the specifications that are currently filled and the information acquired during the dialogue.

Then, the classification GP1 which is the group A, the classification GP2 which is the group B, the classification GPX which is the group X, etc. indicate the options of which group to present.

Then, in FIG. 13, the reward is whether or not the user has purchased the recommended transaction target. That is, the reward is given depending on whether or not the finally recommended transaction target has been purchased. In this way, the information processing apparatus 100 performs reinforcement learning based on the state shown in the state information INF2, the action shown in the classifications GP1, GP2, GPX, etc., and the above reward. For example, the information processing apparatus 100 may select a group to be presented by using a group selection model for selecting a group to be presented from a plurality of groups. In this case, the information processing apparatus 100 may update the group selection model by the reinforcement learning described above. The information processing device 100 is not limited to model-based reinforcement learning, and may perform model-free reinforcement learning and the like.

[1-5-2-1. How to select a group]
The method of selecting the utterance may be appropriately changed depending on the presence or absence of the log. For example, the information processing apparatus 100 selects or searches for groups so that the viewpoints are as different as possible when there is no log of a predetermined amount or more. For example, the information processing apparatus 100 selects a group by using predetermined rule information when there is not enough logs.

Further, for example, the information processing apparatus 100 selects a group having a high expected value of reward when logs are collected. For example, when the information processing apparatus 100 has a predetermined amount or more of logs, the information processing apparatus 100 selects a group having a high expected value of reward. The information processing apparatus 100 may select a group to be presented by using the group selection model described above.

[2. Other embodiments]
The processing according to each of the above-described embodiments may be carried out in various different forms (modifications) other than each of the above-described embodiments. For example, the system configuration is not limited to the above-mentioned example, and may have various aspects. This point will be described below. In the following, the same points as the information processing apparatus 100 according to the embodiment will be omitted as appropriate.

[2-1. Modification example]
For example, in the above example, the information processing device 100, which is a terminal device used by the user, shows an example of performing estimation processing, but information processing that estimates spec values and the like, extracts transaction targets, and generates recommended information. The device and the terminal device used by the user may be separate bodies. This point will be described with reference to FIGS. 14 and 15. FIG. 14 is a diagram showing a configuration example of an information processing system according to a modified example of the present disclosure. FIG. 15 is a diagram showing a configuration example of an information processing device according to a modified example of the present disclosure.

As shown in FIG. 14, the information processing system 1 includes a terminal device 10 and an information processing device 100A. The terminal device 10 and the information processing device 100A are connected to each other via a communication network N so as to be communicable by wire or wirelessly. The information processing system 1 shown in FIG. 14 may include a plurality of terminal devices 10 and a plurality of information processing devices 100A. In this case, the information processing device 100A communicates with the terminal device 10 via the communication network N, provides information to the terminal device 10, and targets character information input by the user via the terminal device 10 as a value of specifications. Etc., extraction of transaction targets, and generation of recommended information may be performed. Further, the information processing device 100A may learn the model based on information such as parameters specified by the user via the terminal device 10.

The terminal device 10 is an information processing device used by the user. The terminal device 10 is realized by, for example, a notebook PC (Personal Computer), a desktop PC, a smartphone, a tablet terminal, a mobile phone, a PDA (Personal Digital Assistant), or the like. The terminal device 10 may be any terminal device as long as it can display the information provided by the information processing device 100A. The terminal device 10 is a client terminal.

The terminal device 10 accepts an utterance by the user as an input. In addition, the terminal device 10 accepts operations by the user. In the example shown in FIG. 14, the terminal device 10 displays the information provided by the information processing device 100A on the screen. Further, the terminal device 10 transmits character information indicating a use case specified by the user to the information processing device 100A. The terminal device 10 transmits character information based on the user's utterance to the information processing device 100A.

The terminal device 10 transmits character information that does not include the value of the corresponding specification to the information processing device 100A. The terminal device 10 transmits character information that does not include a character string indicating specifications to the information processing device 100A. The terminal device 10 transmits character information that does not include a character string indicating a transaction target category to the information processing device 100A. The terminal device 10 transmits character information that does not include a character string indicating a transaction target belonging to the transaction target category to the information processing device 100A.

The terminal device 10 transmits character information indicating the use of the transaction target category to the information processing device 100A. The terminal device 10 transmits character information indicating a usage scene of the transaction target category to the information processing device 100A. The terminal device 10 transmits character information indicating the user's situation to the information processing device 100A. The terminal device 10 transmits character information indicating the usage status of the user with respect to the transaction target of the transaction target category to the information processing device 100A. The terminal device 10 transmits character information indicating whether or not the user is a beginner regarding the use of the transaction target of the transaction target category to the information processing device 100A.

The terminal device 10 displays the information received from the information processing device 100A. The terminal device 10 receives recommended information from the information processing device 100A. The terminal device 10 displays the recommended information received from the information processing device 100A. The terminal device 10 displays recommended information indicating the classification result of classifying the relevant transaction target. The terminal device 10 classifies the relevant transaction target based on the similarity of the relevant transaction target or the purchase history of the relevant transaction target by a similar user similar to the user, and displays recommended information indicating the classification result.

The information processing device 100A realizes the same information processing as the information processing device 100 except that the information processing device 100A is different from the information processing device 100 in that it provides information to the terminal device 10 and acquires information from the terminal device 10. .. The information processing device 100A is a server that provides a service to the terminal device 10 which is a client terminal. For example, the information processing device 100A executes a process of estimating a spec value or the like, extracting a transaction target, and generating recommended information based on the character information acquired from the terminal device 10, and transmits the execution result to the terminal device 10. Send.

As shown in FIG. 15, the information processing device 100A includes a communication unit 11, a storage unit 14, and a control unit 15A. The communication unit 11 is connected to the communication network N (Internet or the like) by wire or wirelessly, and transmits / receives information to / from the terminal device 10 via the communication network N. In this case, the information processing device 100A does not have to have a function of displaying information like the information processing device 100. The information processing device 100A may have an input unit (for example, a keyboard, a mouse, etc.) and a display unit (for example, a liquid crystal display, etc.) used by the administrator of the information processing device 100A.

The control unit 15A is realized by, for example, a CPU, an MPU, or the like executing a program stored in the information processing device 100A (for example, an information processing program according to the present disclosure) with a RAM or the like as a work area. Further, the control unit 15A may be realized by an integrated circuit such as an ASIC or FPGA.

As shown in FIG. 15, the control unit 15A includes an acquisition unit 151A, a learning unit 152, a dialogue management unit 153, an estimation unit 154, an extraction unit 155, a generation unit 156, and a transmission unit 157A. , Realize or execute the information processing functions and actions described below. The internal configuration of the control unit 15A is not limited to the configuration shown in FIG. 15, and may be another configuration as long as it is a configuration for performing information processing described later.

The acquisition unit 151A acquires various information in the same manner as the acquisition unit 151. The acquisition unit 151A acquires various information from the terminal device 10. The acquisition unit 151A acquires user input information from the terminal device 10. The acquisition unit 151A acquires various information from the storage unit 14.

The transmission unit 157A provides various information in the same manner as the transmission unit 157. The transmission unit 157A provides various information to the terminal device 10. The transmission unit 157A transmits various information to the terminal device 10. The transmission unit 157A provides the terminal device 10 with the information generated by the generation unit 156. The transmission unit 157A provides the terminal device 10 with the analysis result by the estimation unit 154. The transmission unit 157A transmits the information to be displayed on the terminal device 10 to the terminal device 10. The transmission unit 157A transmits the recommended information generated by the generation unit 156 to the terminal device 10.

[2-2. Other configuration examples]
In addition, the processing related to each of the above-described embodiments and modifications may be performed in various different forms (modifications) other than the above-described embodiments and modifications. For example, a device that learns a model (learning device), a device that estimates spec values using a model (estimation device), a device that extracts transaction targets (extraction device), and a device that generates recommended information (devices that generate recommended information). The generator) may be a separate body. In this case, the information processing system may include a learning device, an estimation device that is an information processing device that estimates values of specifications and the like, an extraction device, and a generation device. The above is an example, and the information processing system may be realized by various configurations.

[2-3. others]
Further, among the processes described in each of the above embodiments, all or a part of the processes described as being automatically performed can be manually performed, or the processes described as being manually performed. It is also possible to automatically perform all or part of the above by a known method. In addition, the processing procedure, specific name, and information including various data and parameters shown in the above document and drawings can be arbitrarily changed unless otherwise specified. For example, the various information shown in each figure is not limited to the illustrated information.

Further, 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. That is, the specific form of distribution / integration of each device is not limited to the one shown in the figure, and all or part of the device is functionally or physically dispersed / physically distributed in any unit according to various loads and usage conditions. Can be integrated and configured.

Further, each of the above-described embodiments and modifications can be appropriately combined as long as the processing contents do not contradict each other.

Further, the effects described in the present specification are merely examples and are not limited, and other effects may be obtained.

[3. Effect of this disclosure]
As described above, the information processing devices (

information processing devices

100 and 100A in the embodiment) according to the present disclosure include an acquisition unit (acquisition units 151 and 151A in the embodiment) and an estimation unit (estimation unit 154 in the embodiment). To be equipped. The acquisition unit acquires character information indicating a use case specified by the user. The estimation unit estimates the recommended value of the corresponding spec corresponding to the use case among the specifications of the transaction target category corresponding to the use case based on the character information.

As a result, the information processing apparatus according to the present disclosure responds to the use case by the user by estimating the recommended value of the corresponding specifications corresponding to the use case based on the character information indicating the use case specified by the user. The value of the spec can be estimated appropriately. In addition, the information processing device can estimate the value of the spec according to the use case, so that the user can recommend the transaction target of the spec desired by the user without having the user input the value of the spec. ..

In addition, the acquisition unit acquires character information based on the user's utterance. As a result, the information processing device can acquire character information indicating a use case in a dialogue with the user, and can appropriately estimate the value of the specifications according to the use case by the user.

In addition, the estimation unit estimates the recommended value using a model that outputs a score indicating the recommended value in response to the input of character information. As a result, the information processing apparatus can appropriately estimate the value of the specifications according to the use case by the user by using the model that outputs the score indicating the recommended value.

In addition, the estimation unit estimates the recommended value of the spec that the model outputs the score by using the model that outputs a plurality of scores corresponding to each of the plurality of specifications. As a result, the information processing apparatus can appropriately estimate the value for the specifications for which the model outputs the scores by using the model that outputs a plurality of scores corresponding to each of the plurality of specifications.

In addition, the acquisition unit acquires character information that does not include the values of the corresponding specifications. The estimation unit estimates the recommended value based on the character information that does not include the value of the corresponding spec. As a result, the information processing apparatus can appropriately estimate the value of the spec not included in the character information for the preceding character information that does not include the value of the corresponding spec.

In addition, the acquisition unit acquires character information that does not include a character string indicating the specifications. The estimation unit estimates the corresponding specifications for which the recommended value is estimated, based on the character information that does not include the character string indicating the specifications. As a result, the information processing apparatus can appropriately estimate the specifications for which the recommended value should be estimated for the preceding character information that does not include the character string indicating the specifications.

In addition, the acquisition unit acquires character information that does not include a character string indicating a transaction target category. The estimation unit estimates the transaction target category for which the recommended value is estimated, based on the character information that does not include the character string indicating the transaction target category. As a result, the information processing apparatus can appropriately estimate the transaction target category for which the recommended value should be estimated, targeting the preceding character information that does not include the character string indicating the transaction target category.

In addition, the acquisition unit acquires character information that does not include a character string indicating a transaction target that belongs to the transaction target category. The estimation unit estimates the transaction target category for which the recommended value is to be estimated, based on the character string character information indicating the transaction target belonging to the transaction target category. As a result, the information processing apparatus can appropriately estimate the transaction target category for which the recommended value should be estimated, targeting the preceding character information that does not include the character string indicating the transaction target belonging to the transaction target category.

In addition, the acquisition department acquires character information indicating the purpose of the transaction target of the transaction target category. The estimation unit estimates the recommended value of the corresponding specifications of the transaction target category used in the application. As a result, the information processing apparatus can appropriately estimate the value of the specifications according to the use of the transaction target of the transaction target category by the user.

In addition, the acquisition department acquires character information indicating the usage scene of the transaction target of the transaction target category. The estimation unit estimates the recommended value of the corresponding specifications of the transaction target category suitable for use in the usage scene. As a result, the information processing device can appropriately estimate the value of the specifications according to the usage scene of the transaction target of the transaction target category by the user.

In addition, the acquisition unit acquires character information indicating the user's situation. The estimation unit estimates the recommended value of the corresponding specifications of the transaction target category corresponding to the user's situation. As a result, the information processing apparatus can appropriately estimate the value of the specifications according to the user's situation.

In addition, the acquisition unit acquires character information indicating the usage status of the user for the transaction target of the transaction target category. The estimation unit estimates the recommended value of the corresponding specifications of the transaction target category corresponding to the usage status of the user. As a result, the information processing apparatus can appropriately estimate the value of the specifications according to the usage status of the user for the transaction target of the transaction target category.

In addition, the acquisition unit acquires character information indicating whether or not the user is a beginner regarding the use of the transaction target in the transaction target category. When the user is a beginner, the estimation unit estimates the recommended value of the corresponding specifications of the transaction target category to the value corresponding to the beginner. As a result, the information processing apparatus can appropriately estimate the value of the specifications corresponding to the user who is a beginner of the transaction target of the transaction target category.

Further, the information processing apparatus according to the present disclosure includes an extraction unit (in the embodiment, an extraction unit 155). The extraction unit extracts the transaction target whose corresponding spec value corresponds to the recommended value from the transaction target group of the transaction target category as the relevant transaction target. In this way, the information processing device can extract the transaction target suitable for the user's use case by extracting the relevant transaction target from the transaction target group of the transaction target category using the estimated spec value. ..

In addition, the estimation unit estimates a plurality of recommended values corresponding to each of the plurality of specifications. The extraction unit extracts the corresponding transaction target in which each of the plurality of values of the plurality of specifications corresponds to the plurality of recommended values. As a result, the information processing apparatus can extract the transaction target suitable for the user's use case by extracting the corresponding transaction target using a plurality of recommended values corresponding to each of the estimated plurality of specifications. ..

Further, the information processing apparatus according to the present disclosure includes a generation unit (generation unit 156 in the embodiment). The generation unit generates recommended information that recommends the user to purchase the transaction target based on the corresponding transaction target extracted by the extraction unit. In this way, the information processing device can generate recommended information that recommends the purchase of the transaction target to the user based on the relevant transaction target, thereby encouraging the purchase of the transaction target suitable for the user's use case. Become.

In addition, the generation unit generates recommended information when the number of applicable transaction targets is less than the threshold value. As a result, the information processing device can narrow down the transaction target recommended by the user and then promote the purchase of the transaction target suitable for the user's use case. Therefore, the information processing device can increase the possibility that the recommended transaction target is purchased by the user when the recommended information is provided to the user.

In addition, the generation unit generates recommended information indicating the classification result of classifying the relevant transaction target. As a result, the information processing device can classify the transaction target recommended by the user and recommend it to the user. Therefore, when the recommended information is provided to the user, the information processing device makes it easier for the user to select the transaction target by referring to the classification result, and increases the possibility that the recommended transaction target is purchased by the user. be able to.

In addition, the generation unit classifies the relevant transaction target based on the similarity of the relevant transaction target or the purchase history of the relevant transaction target by a similar user similar to the user, and generates recommended information indicating the classification result. As a result, the information processing device can classify the transaction target recommended by the user based on the similarity of the relevant transaction target and the purchase tendency of the similar user, and then recommend the information processing device to the user. Therefore, when the recommended information is provided to the user, the information processing device makes it easier for the user to select the transaction target by referring to the classification result, and increases the possibility that the recommended transaction target is purchased by the user. be able to.

[4. Hardware configuration]
The information devices such as the

information processing devices

100 and 100A according to the above-described embodiments are realized by, for example, a computer 1000 having a configuration as shown in FIG. FIG. 16 is a hardware configuration diagram showing an example of a computer 1000 that realizes the functions of information processing devices such as the

information processing devices

100 and 100A. Hereinafter, the information processing apparatus 100 according to the embodiment will be described as an example. The computer 1000 includes a CPU 1100, a RAM 1200, a ROM (Read Only Memory) 1300, an HDD (Hard Disk Drive) 1400, a communication interface 1500, and an input / output interface 1600. Each part of the computer 1000 is connected by a bus 1050.

The CPU 1100 operates based on the program stored in the ROM 1300 or the HDD 1400, and controls each part. For example, the CPU 1100 expands the program stored in the ROM 1300 or the HDD 1400 into the RAM 1200 and executes processing corresponding to various programs.

The ROM 1300 stores a boot program such as a BIOS (Basic Input Output System) executed by the CPU 1100 when the computer 1000 is started, a program that depends on the hardware of the computer 1000, and the like.

The HDD 1400 is a computer-readable recording medium that non-temporarily records a program executed by the CPU 1100 and data used by the program. Specifically, the HDD 1400 is a recording medium for recording an information processing program according to the present disclosure, which is an example of program data 1450.

The communication interface 1500 is an interface for the computer 1000 to connect to an external network 1550 (for example, the Internet). For example, the CPU 1100 receives data from another device or transmits data generated by the CPU 1100 to another device via the communication interface 1500.

The input / output interface 1600 is an interface for connecting the input / output device 1650 and the computer 1000. For example, the CPU 1100 receives data from an input device such as a keyboard or mouse via the input / output interface 1600. Further, the CPU 1100 transmits data to an output device such as a display, a speaker, or a printer via the input / output interface 1600. Further, the input / output interface 1600 may function as a media interface for reading a program or the like recorded on a predetermined recording medium (media). The media is, for example, an optical recording medium such as DVD (Digital Versatile Disc) or PD (Phase change rewritable Disk), a magneto-optical recording medium such as MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. Is.

For example, when the computer 1000 functions as the information processing device 100 according to the embodiment, the CPU 1100 of the computer 1000 realizes the functions of the control unit 15 and the like by executing the information processing program loaded on the RAM 1200. Further, the information processing program according to the present disclosure and the data in the storage unit 14 are stored in the HDD 1400. The CPU 1100 reads the program data 1450 from the HDD 1400 and executes the program, but as another example, these programs may be acquired from another device via the external network 1550.

The present technology can also have the following configurations.
(1)
An acquisition unit that acquires character information indicating a use case specified by the user,
An estimation unit that estimates the recommended value of the corresponding spec corresponding to the use case among the specifications of the transaction target category corresponding to the use case based on the character information.
Information processing device equipped with.
(2)
The acquisition unit
Acquire the character information based on the utterance of the user,
The information processing device according to (1).
(3)
The estimation unit
The recommended value is estimated using a model that outputs a score indicating the recommended value in response to the input of the character information.
The information processing device according to (1) or (2).
(4)
The estimation unit
Using the model that outputs a plurality of scores corresponding to each of the plurality of specifications, the recommended value of the specifications that the model outputs scores is estimated.
The information processing device according to (3).
(5)
The acquisition unit
Acquire the character information that does not include the value of the corresponding spec, and
The estimation unit
The recommended value is estimated based on the character information that does not include the value of the corresponding specification.
The information processing device according to any one of (1) to (4).
(6)
The acquisition unit
Acquire the character information that does not include the character string indicating the specifications, and
The estimation unit
Based on the character information that does not include the character string indicating the spec, the corresponding spec that is the target of estimation of the recommended value is estimated.
The information processing device according to (5).
(7)
The acquisition unit
Acquire the character information that does not include the character string indicating the transaction target category, and obtain the character information.
The estimation unit
The transaction target category to be the estimation target of the recommended value is estimated based on the character information that does not include the character string indicating the transaction target category.
The information processing device according to (5) or (6).
(8)
The acquisition unit
Acquire the character information that does not include the character string indicating the transaction target belonging to the transaction target category, and obtain the character information.
The estimation unit
A character string indicating a transaction target belonging to the transaction target category The transaction target category to be estimated for the recommended value is estimated based on the character information.
The information processing device according to any one of (5) to (7).
(9)
The acquisition unit
Acquire the character information indicating the use of the transaction target of the transaction target category, and obtain the character information.
The estimation unit
Estimate the recommended value of the corresponding spec of the transaction target category used in the application.
The information processing device according to any one of (1) to (8).
(10)
The acquisition unit
Acquire the character information indicating the usage scene of the transaction target of the transaction target category, and obtain the character information.
The estimation unit
Estimate the recommended value of the corresponding spec of the transaction target category suitable for use in the usage scene.
The information processing device according to (9).
(11)
The acquisition unit
The character information indicating the situation of the user is acquired, and the character information is obtained.
The estimation unit
Estimate the recommended value of the corresponding spec of the transaction target category corresponding to the user's situation.
The information processing device according to any one of (1) to (10).
(12)
The acquisition unit
Acquire the character information indicating the usage status of the user with respect to the transaction target of the transaction target category, and obtain the character information.
The estimation unit
Estimate the recommended value of the corresponding spec of the transaction target category corresponding to the usage situation of the user.
The information processing device according to (11).
(13)
The acquisition unit
Acquire the character information indicating whether or not the user is a beginner regarding the use of the transaction target of the transaction target category.
The estimation unit
When the user is the beginner, the recommended value of the corresponding spec of the transaction target category is estimated to be a value corresponding to the beginner.
The information processing device according to (12).
(14)
An extraction unit that extracts a transaction target whose corresponding spec value corresponds to the recommended value from the transaction target group of the transaction target category as the relevant transaction target.
The information processing apparatus according to any one of (1) to (13).
(15)
The estimation unit
Estimate multiple recommended values corresponding to each of multiple specifications,
The extraction unit
Extract the relevant transaction target in which each of the plurality of values of the plurality of specifications corresponds to the plurality of recommended values.
The information processing device according to (14).
(16)
A generation unit that generates recommended information for recommending the purchase of a transaction target to the user based on the relevant transaction target extracted by the extraction unit.
The information processing apparatus according to (14) or (15).
(17)
The generator
If the number of applicable transaction targets is less than or equal to the threshold value, the recommended information is generated.
The information processing device according to (16).
(18)
The generator
Generate the recommended information showing the classification result of classifying the relevant transaction target.
The information processing apparatus according to (16) or (17).
(19)
The generator
Based on the similarity of the relevant transaction target or the purchase history of the relevant transaction target by a similar user similar to the user, the relevant transaction target is classified and the recommended information indicating the classification result is generated.
The information processing device according to (18).
(20)
Acquires the character information indicating the use case specified by the user,
Based on the character information, among the specifications of the transaction target category corresponding to the use case, the recommended value of the corresponding specifications corresponding to the use case is estimated.
An information processing method that executes processing.

100, 100A Information processing device 11 Communication unit 12 Input unit 13 Audio output unit (speaker)
14 Storage unit 141 Case information storage unit 142 Model information storage unit 143 Transaction target information storage unit 144 Spec list Information storage unit 15, 15A Control unit 151, 151A Acquisition unit 152 Learning unit 153 Dialogue management unit 154 Estimating unit 155 Extraction unit 156

Generation Unit

157, 157A Transmission unit 16 Display unit (display)

Claims

An acquisition unit that acquires character information indicating a use case specified by the user,
An estimation unit that estimates the recommended value of the corresponding spec corresponding to the use case among the specifications of the transaction target category corresponding to the use case based on the character information.
Information processing device equipped with.
The acquisition unit
Acquire the character information based on the utterance of the user,
The information processing device according to claim 1.
The estimation unit
The recommended value is estimated using a model that outputs a score indicating the recommended value in response to the input of the character information.
The information processing device according to claim 1.
The estimation unit
Using the model that outputs a plurality of scores corresponding to each of the plurality of specifications, the recommended value of the specifications that the model outputs scores is estimated.
The information processing device according to claim 3.
The acquisition unit
Acquire the character information that does not include the value of the corresponding spec, and
The estimation unit
The recommended value is estimated based on the character information that does not include the value of the corresponding specification.
The information processing device according to claim 1.
The acquisition unit
Acquire the character information that does not include the character string indicating the specifications, and
The estimation unit
Based on the character information that does not include the character string indicating the spec, the corresponding spec that is the target of estimation of the recommended value is estimated.
The information processing device according to claim 5.
The acquisition unit
Acquire the character information that does not include the character string indicating the transaction target category, and obtain the character information.
The estimation unit
The transaction target category to be the estimation target of the recommended value is estimated based on the character information that does not include the character string indicating the transaction target category.
The information processing device according to claim 5.
The acquisition unit
Acquire the character information that does not include the character string indicating the transaction target belonging to the transaction target category, and obtain the character information.
The estimation unit
A character string indicating a transaction target belonging to the transaction target category The transaction target category to be estimated for the recommended value is estimated based on the character information.
The information processing device according to claim 5.
The acquisition unit
Acquire the character information indicating the use of the transaction target of the transaction target category, and obtain the character information.
The estimation unit
Estimate the recommended value of the corresponding spec of the transaction target category used in the application.
The information processing device according to claim 1.
The acquisition unit
Acquire the character information indicating the usage scene of the transaction target of the transaction target category, and obtain the character information.
The estimation unit
Estimate the recommended value of the corresponding spec of the transaction target category suitable for use in the usage scene.
The information processing device according to claim 9.
The acquisition unit
The character information indicating the situation of the user is acquired, and the character information is obtained.
The estimation unit
Estimate the recommended value of the corresponding spec of the transaction target category corresponding to the user's situation.
The information processing device according to claim 1.
The acquisition unit
Acquire the character information indicating the usage status of the user with respect to the transaction target of the transaction target category, and obtain the character information.
The estimation unit
Estimate the recommended value of the corresponding spec of the transaction target category corresponding to the usage situation of the user.
The information processing device according to claim 11.
The acquisition unit
Acquire the character information indicating whether or not the user is a beginner regarding the use of the transaction target of the transaction target category.
The estimation unit
When the user is the beginner, the recommended value of the corresponding spec of the transaction target category is estimated to be a value corresponding to the beginner.
The information processing device according to claim 12.
An extraction unit that extracts a transaction target whose corresponding spec value corresponds to the recommended value from the transaction target group of the transaction target category as the relevant transaction target.
The information processing apparatus according to claim 1.
The estimation unit
Estimate multiple recommended values corresponding to each of multiple specifications,
The extraction unit
Extract the relevant transaction target in which each of the plurality of values of the plurality of specifications corresponds to the plurality of recommended values.
The information processing device according to claim 14.
A generation unit that generates recommended information for recommending the purchase of a transaction target to the user based on the relevant transaction target extracted by the extraction unit.
The information processing apparatus according to claim 14.
The generator
If the number of applicable transaction targets is less than or equal to the threshold value, the recommended information is generated.
The information processing device according to claim 16.
The generator
Generate the recommended information showing the classification result of classifying the relevant transaction target.
The information processing device according to claim 16.
The generator
Based on the similarity of the relevant transaction target or the purchase history of the relevant transaction target by a similar user similar to the user, the relevant transaction target is classified and the recommended information indicating the classification result is generated.
The information processing device according to claim 18.
Acquires the character information indicating the use case specified by the user,
Based on the character information, among the specifications of the transaction target category corresponding to the use case, the recommended value of the corresponding specifications corresponding to the use case is estimated.
An information processing method that executes processing.