Interaction device and program

Abstract

To provide an interaction device capable of performing an appropriate interaction with a user.SOLUTION: An interaction device 1 that performs an interaction with a user comprises: a question generation unit 72 that generates questions; an evaluation unit 73 that evaluates the user's response to the question; and an interaction development unit 74 that develops the interaction on the basis of the evaluation. The evaluation unit 73 evaluates whether or not the response includes all necessary items that should be answered in response to the question. The interaction development unit 74 generates a question regarding the missing item when at least a portion of the necessary items is not included in the response, and generates a new question related to the response when all the necessary items are included in the response.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a dialogue device and a program for conducting a dialogue with a user.

Background Art

[0002] In English conversation classes, pre-learning is very important. If pre-learning to a certain extent is not carried out in advance and the learner cannot have basic conversations, in the actual class, the conversation with the instructor will not go well, and it will be difficult to enhance the effect of the class.

[0003] Therefore, English dialogue apps for assisting pre-learning have been developed. For example, Non-Patent Documents 1 and 2 disclose apps that use artificial intelligence (AI) to conduct a dialogue with a user.

Prior Art Documents

Non-Patent Documents

[0004]

Non-Patent Document 1

Non-Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, different from a program based on rules, the dialogue by artificial intelligence is difficult to predict how it will develop, and there is a possibility that the development will be in conversation sentences different from what the learner intends to learn, and it is assumed that the learning will be inappropriate for the learner.

[0006] The present invention has been made in view of the above problems, and an object thereof is to provide a dialogue device capable of appropriate dialogue with a user.

Means for Solving the Problems

[0007] In order to solve the above problems, the present invention includes the following aspects. Item 1. A dialogue device for conducting a dialogue with a user, a question generation unit that generates questions, an evaluation unit that evaluates the user's response to the question, a dialogue development unit that develops a dialogue based on the evaluation, and comprising the evaluation unit evaluates whether all of the necessary items to answer the question are included in the response, when at least a part of the necessary items is not included in the response, the dialogue development unit generates a question regarding the item not included, a dialogue device that generates a new question regarding the response when all of the necessary items are included in the response. Item 2. The dialogue according to Item 1, wherein the dialogue is conducted in a foreign language that the user is learning. Item 3. Each function of the question generation unit, the evaluation unit, and the dialogue development unit is realized by using a machine-learned artificial intelligence model, and the dialogue device according to Item 1 or 2. Item 4. A program for causing a computer to function as each part of the dialogue device according to any one of Items 1 to 3.

Effects of the Invention

[0008] According to the present invention, it is possible to provide a dialogue device capable of appropriate dialogue with a user.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Embodiments for Carrying Out the Invention

[0010] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. Note that the present invention is not limited to the following embodiments, and various modifications are possible without departing from the spirit thereof.

[0011] (Overall Configuration) FIG. 1 is a block diagram showing the configuration of a dialogue device 1 according to an embodiment of the present invention. The dialogue device 1 is a device that conducts a dialogue in a foreign language with a user learning the foreign language, and can be configured by a smartphone, a general-purpose personal computer, or the like. In the present embodiment, the foreign language is English, but the type of language is not particularly limited. Further, in the present embodiment, the dialogue device 1 is mainly used for a user to perform pre-dialogue practice before a face-to-face class with an instructor.

[0012] The dialogue device 1 includes a microphone 2, a speaker 3, an input unit 4, a display unit 5, a character / audio conversion unit 6, and a dialogue control unit 7.

[0013] The microphone 2 and the speaker 3 may be built into the dialogue device 1 or externally attached to the dialogue device 1.

[0014] The input unit 4 is a device for a user to operate on the dialogue device 1, and can be configured by a touch panel, a keyboard, a mouse, or the like.

[0015] The display unit 5 can be configured by a liquid crystal display or an organic EL display.

[0016] The text / audio conversion unit 6 converts the audio data collected by the microphone 2 into text data and outputs it to the dialogue control unit 7, and also converts the text data input from the dialogue control unit 7 into audio data and outputs it to the speaker 3. The function of the text / audio conversion unit 6 can be realized by using a known application or the like.

[0017] The dialogue control unit 7 has a function of controlling the dialogue between the user and the dialogue device 1. The dialogue control unit 7 can be realized software-wise by the processor of the dialogue device 1 executing the program P stored in the storage 8. The program P may be downloaded to the dialogue device 1 via a communication network such as the Internet, or the program P may be recorded on a computer-readable non-transitory recording medium such as a CD-ROM or an SD card, and installed on the dialogue device 1 via the storage medium.

[0018] The program P may be a rule-based program, or the function of controlling the dialogue between the user and the dialogue device 1 may be realized in the dialogue device 1 by using a machine-learned AI model. In the present embodiment, the functions of the question generation unit 72, the evaluation unit 73, and the dialogue development unit 74 of the dialogue control unit 7 are realized by using the machine-learned dialogue AI model M. The dialogue AI model M is stored in the server 9 on the cloud and is available to the dialogue control unit 7 through API linkage, but it may also be stored in the dialogue device 1. Examples of the dialogue AI model M include AI chatbots such as ChatGPT.

[0019] (Dialogue Control) FIG. 2 is a more specific functional block diagram of the dialogue control unit 7. The dialogue control unit 7 includes a lesson selection unit 71, a question generation unit 72, an evaluation unit 73, and a dialogue development unit 74. The functions of these blocks will be described with reference to the flowchart shown in FIG. 3.

[0020] When the program P is started, in step S1, the course selection unit 71 selects a dialogue lesson in response to a user's operation. In the program P, lessons for dialogue practice corresponding to various levels and topics are prepared, and the user can select a lesson according to the class he or she plans to take. In this embodiment, it is assumed that a beginner-level lesson is selected.

[0021] Next, in step S2, the question generator 72 generates a question for the user according to the selected lesson. The question for the user may be prepared in advance in association with the lesson. Alternatively, the question generator 72 may instruct the dialogue AI model M to generate a question according to the level and topic of the selected lesson, and obtain the question generated by the dialogue AI model M in response to the instruction.

[0022] The question generator 72 generates a question as character data, and the character data is converted by the character / speech converter 6 into voice data, which is then output from the speaker 3. In this embodiment, it is assumed that the following question is generated. Tell me about some of your plans for next week.

[0023] When the user responds to this (YES in step S3), the process proceeds to step S4, where the evaluation unit 73 evaluates the user's response to the question. Specifically, the evaluation unit 73 evaluates whether the response includes all of the necessary items to be answered in response to the question that are expected in the lesson. The necessary items are set for each question. For example, the necessary items to be answered in response to the above question are set to the following three items. (1) Details of the Plan (2) Where to carry out the plan (3) Someone to help you with your plans

[0024] That is, the user is required to answer what, where, and with whom they plan to do. Note that the number of necessary items may be one or more, and the more necessary items can be set as the learning level increases.

[0025] The dialogue development unit 74 shown in FIG. 1 develops the dialogue based on the evaluation by the evaluation unit 73. Specifically, when at least a part of the necessary items is not included in the user's response (NO in step S5), it proceeds to step S6, and when all of the necessary items are included in the user's response (YES in step S5), it proceeds to step S7.

[0026] For example, if the user's response is I'm going to go shopping. (I plan to go shopping.) assuming that, since (2) and (3) of the necessary items are not included in the response, it proceeds to step S6. On the other hand, if the user's response is I'm going to go shopping in Osaka with my friends. (I plan to go shopping in Osaka with my friends.) assuming that, since all of the necessary items (1) to (3) are included in the response, it proceeds to step S7.

[0027] The determination of whether all of the necessary items are included in the response can be made by the dialogue AI model M. Even when not using the dialogue AI model M, it is possible to determine whether all of the necessary items are included in the response by identifying the verb in the response and the noun following the preposition. For example, by identifying Osaka as the noun following in and searching in a database or the like to see if Osaka exists as a word indicating a location, it is possible to determine whether the necessary item (2) is included in the response.

[0028] When transitioning to step S6, the dialogue development unit 74 generates questions regarding matters not included in the response (step S6). For example, if (2) and (3) among the necessary matters are not included in the response, the dialogue development unit 74 generates questions regarding the necessary matters (2) and (3). Specifically, Where are you going to go? (Where do you plan to go?) Who are you going to go with? (Who do you plan to go with?) Questions such as these may be issued to the user one by one, or may be issued to the user collectively.

[0029] Questions regarding the necessary matters can be made to the dialogue AI model M. Specifically, by presenting the necessary matters as slot information to the dialogue AI model M and inputting a prompt instructing it to question the items of the slot information, the dialogue AI model M can automatically generate questions regarding the necessary matters. Even when not using the dialogue AI model M, questions corresponding to the necessary matters (1) to (3) can be stored in advance in the database, and by searching for questions corresponding to matters not included in the response, questions regarding the necessary matters can be generated.

[0030] When the user responds to the questions regarding the necessary matters (YES in step S3), the process returns to step S4 again, and the evaluation unit 73 evaluates the user's response to the questions. Then, step S6 is repeated until the user has responded to all of the necessary matters (YES in step S5).

[0031] When transitioning to step S7, the dialogue development unit 74 generates new questions regarding the response (step S7). New questions regarding the response mean questions regarding matters newly arising from the user's response that were not included in the questions generated in step S2.

[0032] In the example of this embodiment, due to the user's response, · The planned activity is shopping · The place to execute the plan is Osaka · The person to execute the plan together with is a friend These become new matters. The dialogue development unit 74 generates questions (in-depth questions) to elicit more detailed content of the new matters. Examples of such questions include the following questions. What are you going to buy? (What do you plan to buy?) Do you often go there? (Do you go there frequently?) Which shop are you going to go? (Which store do you plan to go to?)

[0033] New questions regarding the response can be given to the dialogue AI model M. Specifically, by inputting a prompt instructing the dialogue AI model M to dig deeper into the above slot information, the dialogue AI model M can automatically generate new questions regarding the response. Even when not using the dialogue AI model M, questions related to the matters included in the response are stored in the database in advance, and new questions regarding the response can be generated by searching for questions corresponding to the matters newly arising from the response.

[0034] After that, when the number of turns of the dialogue reaches a predetermined number (YES in step S8), the dialogue ends. Instead of ending the dialogue, a new dialogue unrelated to the previous topic may be developed by uttering "By the way".

[0035] (Summary) As described above, in the dialogue device 1 according to the present embodiment, the evaluation unit 73 evaluates whether or not all of the necessary items to be answered for the question are included in the user's response to the question, and the dialogue development unit 74 generates a question regarding the item not included when at least a part of the necessary items is not included in the response. In the prior art, even when the user gives an incomplete response, there is a risk that the AI will freely develop the topic and the user will not be able to conduct appropriate dialogue practice. On the other hand, the dialogue device 1 according to the present embodiment can prompt the user to respond until the necessary items to be answered are answered, so that appropriate dialogue with the user becomes possible.

[0036] In the present embodiment, although the AI (dialogue AI model M) is used in the same manner as in the prior art, since the dialogue control unit 7 controls the dialogue AI model M by means of a prompt, inappropriate development of the dialogue is prevented.

Industrial Applicability

[0037] In the above embodiment, the user learning a foreign language uses the dialogue device to perform pre-dialogue practice before face-to-face lessons with an instructor. However, the application of the present invention is not limited to pre-dialogue practice before face-to-face lessons. Further, in the above embodiment, the dialogue is conducted in the foreign language that the user is learning. However, the present invention is not limited to foreign language learning and may be applied to dialogue in the user's mother tongue using a computer.

Explanation of Reference Numerals

[0038] 1 Dialogue device 2 Microphone 3 Speaker 4 Input unit 5 Display unit 6 Character / audio conversion unit 7 Dialogue control unit 71 Course selection unit 72 Question generation unit 73 Evaluation unit 74 Dialogue development unit 8 Storage 9 Server AI model for M conversation (artificial intelligence model) P Program

Claims

Claim 1 An interactive device for interacting with a user, comprising: a question generation unit that generates questions; an evaluation unit that evaluates the user's response to the question; a dialogue development unit that develops a dialogue based on the evaluation; wherein the evaluation unit evaluates whether all of the necessary items to answer the question are included in the response; when at least a part of the necessary items are not included in the response, the dialogue development unit generates a question regarding the items not included; when all of the necessary items are included in the response, the dialogue development unit generates a new question regarding the response, the interactive device. Claim 2 The interactive device according to claim 1, wherein the dialogue is conducted in a foreign language that the user is learning. Claim 3 The interactive device according to claim 1, wherein each function of the question generation unit, the evaluation unit, and the dialogue development unit is realized by using a machine-learned artificial intelligence model. Claim 4 A program that causes a computer to function as each part of the interactive device according to any one of claims 1 to 3.

Images