Interview device, interview method, and program

Abstract

To provide an interview device, interview method and program, which enable efficient and accurate interviews with interviewees.SOLUTION: An interview device disclosed herein comprises: an input unit for inputting a first text representing an answer of an interviewee and an instruction to ask a question to the answer as an input sentence to a sentence generation model configured to generate an output sentence in response to the input sentence; and an output unit for outputting a question to the interviewee as the output sentence generated by the sentence generation model.SELECTED DRAWING: Figure 5

Description

[Technical field]

[0001] The technology disclosed herein relates to an interview device, an interview method, and a program. [Background technology]

[0002] Patent document 1 discloses an interview system in which an interviewer terminal and an interviewee terminal are connected to a conference server via a network, and the interviewer terminal and the interviewee terminal access the conference server using authentication information, thereby enabling an online interview to be conducted using the conference server.The interview system further includes an interview setting server to which the interviewer terminal and the interviewee terminal are connected via the network, and the interview setting server generates interview information including an interview date and time based on access from the interviewee terminal and transmits the interview information to the conference server, and the conference server issues authentication information based on the transmission of the interview information from the interview setting server. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-004142 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technology described in Patent Document 1 leaves room for improvement in terms of conducting interviews with interviewees efficiently and accurately.

[0005] The present disclosure aims to provide an interview device, an interview method, and a program that can efficiently and accurately conduct interviews with interviewees. [Means for solving the problem]

[0006] A first aspect of the technology disclosed herein is an interview device comprising: an input unit that inputs a first text representing an interviewee's answer and an instruction to ask a question in response to the answer as an input sentence into a sentence generation model that generates an output sentence according to an input sentence; and an output unit that outputs a question as the output sentence generated by the sentence generation model to the interviewee.

[0007] A second aspect of the technology disclosed herein is an interview method in which a computer executes a process including inputting a first text representing an interviewee's answer and an instruction to ask a question in response to the answer as an input sentence into a sentence generation model that generates an output sentence in response to the input sentence, and outputting the question as the output sentence generated by the sentence generation model to the interviewee.

[0008] A third aspect of the technology disclosed herein is a program for causing a computer to execute a process including inputting a first text representing an interviewee's answer and an instruction to ask a question in response to the answer as an input sentence into a sentence generation model that generates an output sentence in response to the input sentence, and outputting the question as the output sentence generated by the sentence generation model to the interviewee. [Brief description of the drawings]

[0009] [Figure 1] A block diagram showing an example of a hardware configuration of the interview system. [Diagram 2] A block diagram showing an example of the functional configuration of an interview device. [Diagram 3] FIG. 11 is a first explanatory diagram showing an example of the operation of the interview device. [Figure 4] FIG. 2 is a second explanatory diagram showing an example of the operation of the interview device. [Diagram 5] FIG. 3 is a third explanatory diagram showing an example of the operation of the interview device. [Figure 6] FIG. 4 is a fourth explanatory diagram showing an example of the operation of the interview device. [Figure 7] FIG. 5 is an explanatory diagram showing an example of the operation of the interview device. [Figure 8] FIG. 6 is an explanatory diagram showing an example of the operation of the interview device. [Figure 9] FIG. 7 is an explanatory diagram showing an example of the operation of the interview device. [Figure 10] FIG. 8 is an explanatory diagram showing an example of the operation of the interview device. [Figure 11] FIG. 9 is a ninth explanatory diagram showing an example of the operation of the interview device. [Figure 12] 13 is a flowchart showing an example of the flow of an interview execution process. [Figure 13] FIG. 1 is a conceptual diagram illustrating an example of an emotion map. [Figure 14] FIG. 1 is a conceptual diagram illustrating an example of an emotion map. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Below, an example of an embodiment of an interview device, an interview method, and a program related to the technology disclosed herein will be described with reference to the attached drawings.

[0011] First, the terms used in the following description will be explained.

[0012] In the following embodiments, a signed processor (hereinafter simply referred to as a "processor") may be one arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be one type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), or a TPU (Tensor Processing Unit).

[0013] In the following embodiments, a signed RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0014] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0015] In the following embodiments, a communication I / F (Interface) with a code is an interface including a communication processor and an antenna. The communication I / F controls communication between multiple computers. An example of a communication standard applied to the communication I / F is a wireless communication standard including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0016] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. In addition, in this specification, the same idea as "A and / or B" is also applied when three or more things are expressed by connecting them with "and / or."

[0017] FIG. 1 shows an example of the configuration of an interview system 10 according to the embodiment.

[0018] As shown in FIG. 1, the interview system 10 comprises an interview device 12 and an interviewee terminal 14. The interview device 12 is a device that functions as an AI interviewer, and the interviewee terminal 14 is a device that functions as an interface with the interviewee. An example of the interview device 12 is a server. An example of the interviewee terminal 14 is a notebook personal computer, a smartphone, or a tablet terminal. Below, a case where the interviewee terminal 14 is a notebook personal computer will be described.

[0019] The interview device 12 comprises a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 comprises a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network), etc.

[0020] The interviewee terminal 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0021] The reception device 38 includes a keyboard 38A, a mouse 38B, a touchpad 38C, a microphone 38D, etc., and receives input from the interviewee. When the reception device 38 receives input from the interviewee, it outputs data corresponding to the input to the processor 46.

[0022] The output device 40 includes a display 40A and a speaker 40B, and presents data to the person 20 by outputting the data in a form of expression that the person 20 can perceive (e.g., voice and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs voice according to instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a Complementary Metal-Oxide-Semiconductor (CMOS) image sensor or a Charge Coupled Device (CCD) image sensor.

[0023] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54.

[0024] FIG. 2 shows an example of the main functions of the interview device 12.

[0025] As shown in Figure 2, in the interview device 12, interview execution processing is performed by the processor 28. An interview execution program 56 is stored in the storage 32. The interview execution program 56 is an example of a "program" related to the technology of the present disclosure. The processor 28 reads the interview execution program 56 from the storage 32, and executes the read interview execution program 56 on the RAM 30. The interview execution processing is realized by the processor 28 operating as a question start unit 28A, an input unit 28B, and an output unit 28C in accordance with the interview execution program 56 executed by the processor 28 on the RAM 30.

[0026] The storage 32 stores a sentence generation model 58 and an emotion identification model 60. The sentence generation model 58 and the emotion identification model 60 are used in the interview execution process. The sentence generation model 58 is a trained model that generates output sentences according to input sentences. Specifically, the sentence generation model 58 is a so-called generative AI (Artificial Intelligence). One example of the sentence generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The sentence generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the sentence generation model 58, and inference data such as voice data indicating a voice, text data indicating a text, and image data indicating an image are input. The sentence generation model 58 infers the input inference data according to the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. Although FIG. 2 shows an embodiment in which the sentence generation model is stored in a storage, the sentence generation model may be stored in a cloud computer communicably connected to the interview device. The emotion identification model 60 will be described later.

[0027] Next, with reference to Figures 3 to 11, an example of the operation of the question initiation unit 28A, the input unit 28B, and the output unit 28C when the interview device 12 performs an interview execution process to conduct an interview with an interviewee will be described.

[0028] As shown in FIG. 3, when the interview starts, the question initiation unit 28A outputs the first question, "Question 1," to the interviewee terminal 14. In the example shown in FIG. 3, "Question 1" is "First, please tell me your name, educational background, and motivation for applying." The first question, "Question 1," may be a predetermined question. The interviewee terminal 14 outputs "Question 1" to the interviewee. The interviewee terminal 14 may output "Question 1" by voice or by text. Furthermore, the interviewee terminal 14 may output "Question 1" by both voice and text. The interviewee terminal 14 may also output each of the following questions by voice, by text, or by both voice and text.

[0029] The interviewee inputs "Answer 1" to the interviewee terminal 14 as an answer to "Question 1". In the example shown in FIG. 3, "Answer 1" is "My name is XX. My academic background is XX. I am applying to your company because I am strongly motivated to work for a world-leading company in XX." The interviewee may input "Answer 1" by voice or in text. The interviewee may also input "Answer 1" by both voice and text. The interviewee may also input each of the following answers by voice, in text, or in both voice and text. The interviewee terminal 14 outputs "Answer 1" to the interview device 12.

[0030] The input unit 28B has a function of determining a policy for proceeding with the interview based on the answers of the interviewee and / or the evaluation items for the interviewee. The input unit 28B may, for example, extract terms included in the answers and determine a policy corresponding to the extracted terms using pattern matching, artificial intelligence, or the like. The input unit 28B may also, for example, acquire evaluation items set in advance by the interviewer and determine a policy corresponding to the acquired evaluation items using pattern matching, artificial intelligence, or the like. The evaluation items are items that the interviewer wants to include in the interview summary and can be set arbitrarily by the interviewer.

[0031] The evaluation items may include, for example, at least one of the interviewee's humanity, enthusiasm, and sociability. Enthusiasm is an item for evaluating the willingness to improve oneself, enthusiasm for work, initiative, etc. The evaluation items for humanity and sociability may include the following items. Emotion: Items to estimate the emotional state (positive, anxious, excited, positive, negative, etc.) at the time of response from tone, voice strength, speech rate, etc. Motivation: Items to evaluate motivation for applying, commitment to the organization, etc. Communication skills: Items to evaluate communication skills from linguistic expressions such as clarity and fluency of responses, vocabulary, and appropriate responses to questions. Leadership and Teamwork: Items designed to assess your experience working in a team and how you would demonstrate leadership in specific scenarios. Responsibility: Items to evaluate whether you can take responsibility for negative results, admit mistakes, and strive to improve Ethics and values: Items to evaluate the candidate's ethics, values, morality, and awareness of norms (compliance) from each individual response. Adaptability: Items to evaluate whether you can adjust your actions and words according to the situation, and whether you can handle multiple tasks simultaneously. Stress tolerance and coping skills: Items to assess stress management and problem-solving skills based on reactions to unexpected questions or difficult situations Logical thinking: Items to evaluate the level of understanding of a given question or situation and the logic of the response to it. Creativity and Innovation: Items to assess whether you can propose new ideas and have an innovative approach to solving problems Problem-solving skills: items to assess your ability to break down complex problems and come up with appropriate solutions Information gathering ability: Items to evaluate whether you can gather information efficiently and act on it. Self-awareness: Items designed to assess whether you understand your own strengths and weaknesses and are motivated to work on self-improvement Decision-making ability: Items to assess whether you can make efficient and effective decisions with limited information. Specialized knowledge and technical skills: Items to evaluate whether you have the specialized knowledge and technical skills required for the job you are applying for.

[0032] The input unit 28B may determine a policy regarding the progress of the interview based on whether or not sufficient information has been obtained and on the elapsed time.

[0033] In the example shown in FIG. 3, input unit 28B determines a policy that "the motivation for applying should be explored further" based on "Answer 1" and the evaluation items set in advance by the interviewer.

[0034] Then, the input unit 28B generates a prompt, which is a text representing "Question 1", "Answer 1", and "An instruction to ask a question about Answer 1", according to the determined policy, and inputs the generated prompt to the sentence generation model 58 as an input sentence. In this way, the input unit 28B has a function as a prompt input unit that inputs a prompt to the sentence generation model 58. In the example shown in FIG. 3, the input unit 28B generates a prompt saying, "The student's answer is 'My name is XX ~ My motivation is.' Please think of the next question for this student." and inputs the generated prompt to the sentence generation model 58 as an input sentence. The text generated in this case is an example of the "first text" in this disclosure.

[0035] In response to the prompt input by the input unit 28B, the sentence generation model 58 generates the next question, "Question 2," as an output sentence. In the example shown in Fig. 3, "Question 2" is "Thank you for your self-introduction. Please tell us if there are any specific projects or areas that you would like to work on at this company."

[0036] The output section 28C outputs "Question 2", which is a question as an output sentence generated by the sentence generation model 58, to the interviewee terminal 14. The interviewee terminal 14 outputs "Question 2" to the interviewee.

[0037] As shown in Fig. 4, the interviewee inputs "Answer 2", which is an answer to "Question 2", into the interviewee terminal 14. The interviewee terminal 14 outputs "Answer 2" to the interview device 12. In the example shown in Fig. 4, "Answer 2" is "I am particularly interested in XX technology, and would like to participate in a project that utilizes that technology to solve a social problem."

[0038] The input unit 28B determines a policy when "Answer 2" is input to the interview device 12. In the example shown in Fig. 4, the input unit 28B determines a policy that "the applicant's motivation for applying should be explored further" based on "Answer 2" and the evaluation items set in advance by the interviewer.

[0039] Then, the input unit 28B generates a prompt, which is a text representing "Question 2", "Answer 2", and "An instruction to ask a question about Answer 2", according to the determined policy, and inputs the generated prompt to the sentence generation model 58 as an input sentence. In the example shown in FIG. 4, the input unit 28B generates a prompt saying, "In response to the question, 'Thank you for your self-introduction, please tell us what you would like to work on at this company,' the student answers, 'I would like to participate in XX technology.' Please think of the following question for this student." and inputs the generated prompt to the sentence generation model 58 as an input sentence. The text generated in this case is an example of the "first text" in this disclosure.

[0040] The sentence generation model 58 generates the next question, "Question 3," as an output sentence in response to the prompt input by the input unit 28B. In the example shown in Fig. 4, "Question 3" is "Could you tell me in detail how you became interested in XX technology and what specific social problem you want to solve?"

[0041] The output section 28C outputs "Question 3", which is a question as an output sentence generated by the sentence generation model 58, to the interviewee terminal 14. The interviewee terminal 14 outputs "Question 3" to the interviewee.

[0042] As shown in FIG. 5, the interviewee inputs "Answer 3," which is an answer to "Question 3," into the interviewee terminal 14. The interviewee terminal 14 outputs "Answer 3" to the interview device 12. In the example shown in FIG. 5, "Answer 3" is, "It all started when I started XX in high school and took a XX class in college. Specifically, I'm interested in a XX support system that utilizes XX. I believe that if we can use AI technology to solve problems that occur in the XX field, we can improve the lives of many people."

[0043] The input unit 28B determines a policy when "Answer 3" is input to the interview device 12. In the example shown in Fig. 5, the input unit 28B determines a policy that "Since information has been acquired, the next question should be asked. The next question should measure 'the ability to handle unexpected situations', which was set in advance by the interviewer," based on "Answer 3" and the evaluation items set in advance by the interviewer.

[0044] Then, the input unit 28B generates a prompt, which is a text representing "Question 3", "Answer 3", and "An instruction to ask a question about Answer 3", according to the determined policy, and inputs the generated prompt to the sentence generation model 58 as an input sentence. In the example shown in FIG. 5, the input unit 28B generates a prompt saying, "In response to the question, 'Are you interested in XX technology?', the student's answer is, 'I'm thinking about XX in high school.' I understand the motivation well, so next I would like to measure the ability to deal with unexpected situations. Please think of a question to measure that ability.'" and inputs the generated prompt to the sentence generation model 58 as an input sentence. The text generated in this case is an example of the "first text" in this disclosure.

[0045] In response to the prompt input by the input unit 28B, the sentence generation model 58 generates the next question, "Question 4," as an output sentence. In the example shown in Fig. 5, "Question 4" is "Who is the person you respect the most? Also, can you tell me why?"

[0046] The output section 28C outputs "Question 4", which is a question as an output sentence generated by the sentence generation model 58, to the interviewee terminal 14. The interviewee terminal 14 outputs "Question 4" to the interviewee.

[0047] As shown in Fig. 6, the interviewee inputs "Answer 4", which is an answer to "Question 4", to the interviewee terminal 14. The interviewee terminal 14 outputs "Answer 4" to the interview device 12. In the example shown in Fig. 6, "Answer 4" is "The person I respect most is XX, the founder of XX. I am deeply impressed by his vision and the passion to turn it into reality."

[0048] The input unit 28B determines a policy when "Answer 4" is input to the interview device 12. In the example shown in Fig. 6, the input unit 28B determines a policy that "There is insufficient information, so we should dig deeper" based on "Answer 4" and the evaluation items set in advance by the interviewer.

[0049] Then, the input unit 28B generates a prompt, which is a text representing "Question 4", "Answer 4", and "An instruction to ask a question about Answer 4", according to the determined policy, and inputs the generated prompt to the sentence generation model 58 as an input sentence. In the example shown in FIG. 6, the input unit 28B generates a prompt saying, "In response to the question, 'Do you most respect ~?', the student's answer is, 'The person I respect most is ~.' In this conversation, please ask a question that will further dig into the situation-handling ability." and inputs the generated prompt to the sentence generation model 58 as an input sentence. The text generated in this case is an example of the "first text" in this disclosure.

[0050] In response to the prompt input by the input unit 28B, the sentence generation model 58 generates the next question, "Question 5," as an output sentence. In the example shown in Fig. 6, "Question 5" is "XX is certainly respected by many people. So, if XX were alive, and you could ask him just one question, what would you ask him?"

[0051] As shown in Fig. 7, the interviewee inputs "Answer 5", which is an answer to "Question 5", to the interviewee terminal 14. The interviewee terminal 14 outputs "Answer 5" to the interview device 12. In the example shown in Fig. 7, "Answer 5" is "If I could ask him a question, I would like to ask him, 'How have you dealt with the difficulties and challenges you have faced?' Because I think that would help me find a way to deal with my own difficulties."

[0052] The input unit 28B determines a policy when "Answer 5" is input to the interview device 12. In the example shown in Fig. 7, the input unit 28B determines a policy that "Since information has been acquired, the next question should be asked. Next, we should ask about the experiences during school days, which were set in advance by the interviewer," based on "Answer 5" and the evaluation items set in advance by the interviewer.

[0053] Then, the input unit 28B generates a prompt, which is a text representing "Question 5", "Answer 5", and "An instruction to ask a question about Answer 5", according to the determined policy, and inputs the generated prompt to the sentence generation model 58 as an input sentence. In the example shown in FIG. 7, the input unit 28B generates a prompt saying, "In response to the question, 'Do you really want to hear about XX?', the student's answer is 'If he...'. Next, please think of a question to ask this student about what he focused on during his student days." and inputs the generated prompt to the sentence generation model 58 as an input sentence. The text generated in this case is an example of the "first text" in this disclosure.

[0054] In response to the prompt input by the input unit 28B, the sentence generation model 58 generates the next question, "Question 6," as an output sentence. In the example shown in Fig. 7, "Question 6" is "Thank you. Next question. What did you focus on most when you were a student?"

[0055] As shown in Figure 8, the interviewee inputs "Answer 6", which is an answer to "Question 6", to the interviewee terminal 14. The interviewee terminal 14 outputs "Answer 6" to the interview device 12. In the example shown in Figure 8, "Answer 6" is "The thing I focused on most when I was a student was XXX."

[0056] The input unit 28B determines a policy when "Answer 6" is input to the interview device 12. In the example shown in Fig. 8, the input unit 28B determines the policy that "Deeper investigation is required" based on "Answer 6" and the evaluation items set in advance by the interviewer.

[0057] Then, the input unit 28B generates a prompt, which is a text representing "Question 6", "Answer 6", and "An instruction to ask a question about Answer 6", according to the determined policy, and inputs the generated prompt to the sentence generation model 58 as an input sentence. In the example shown in FIG. 8, the input unit 28B generates a prompt saying, "In response to the question, 'What did you focus on most when you were a student?', the student answers, 'The thing I focused on most when I was a student was XXX.' Please think of the following question for this student." and inputs the generated prompt to the sentence generation model 58 as an input sentence. The text generated in this case is an example of the "first text" in this disclosure.

[0058] In response to the prompt input by the input unit 28B, the sentence generation model 58 generates the next question, "Question 7," as an output sentence. In the example shown in Fig. 8, "Question 7" is "Do you have any specific stories about XXX?"

[0059] As shown in Figure 9, the interviewee inputs "Answer 7", which is an answer to "Question 7", to the interviewee terminal 14. The interviewee terminal 14 outputs "Answer 7" to the interview device 12. In the example shown in Figure 9, "Answer 7" is "The specific episode is XX."

[0060] The input unit 28B determines a policy when "Answer 7" is input to the interview device 12. In the example shown in Fig. 9, the input unit 28B determines a policy that "I was able to ask what I should have asked in the interview. As the time is approaching the end, we should move on to the final question" based on "Answer 7" and the evaluation items set in advance by the interviewer.

[0061] Then, the input unit 28B generates a prompt, which is a text representing "Question 7", "Answer 7", and "An instruction to ask a question about Answer 7", according to the determined policy, and inputs the generated prompt as an input sentence to the sentence generation model 58. In the example shown in FIG. 9, the input unit 28B generates a prompt saying, "In response to the question, 'Do you have a specific episode about XXX?', the student's answer is, 'The specific episode is XXX.' Please think about the last question." and inputs the generated prompt as an input sentence to the sentence generation model 58. The text generated in this case is an example of the "first text" in this disclosure.

[0062] The sentence generation model 58 generates the next question, "Question 8," as an output sentence in response to the prompt input by the input unit 28B. In the example shown in Fig. 9, "Question 8" is "Do you have any final questions? We will answer them by email at a later date."

[0063] As shown in Figure 10, the interviewee inputs "Answer 8", which is an answer to "Question 8", to the interviewee terminal 14. The interviewee terminal 14 outputs "Answer 8" to the interview device 12. In the example shown in Figure 10, "Answer 8" is "Regarding the company's system, is XX XX?"

[0064] When "Answer 8" is input to the interview device 12, the input unit 28B generates a prompt, which is a text representing "Question 8", "Answer 8", and "An instruction to ask a question about Answer 8", according to the determined policy, and inputs the generated prompt to the sentence generation model 58 as an input sentence. In the example shown in FIG. 10, since the end condition of the interview is satisfied, the input unit 28B generates a prompt saying "In response to the question, 'Do you have any final questions? We will answer you by email at a later date,' the student answers, 'Is it the company's system?' Please give this student a closing remark." and inputs the generated prompt to the sentence generation model 58 as an input sentence. The text generated in this case is an example of the "first text" in this disclosure.

[0065] The sentence generation model 58 generates the next question, "Question 9," as an output sentence in response to the prompt input by the input unit 28B. In the example shown in Fig. 10, "Question 9" is "Thank you for your question. A person in charge will respond to you at a later date. Thank you for your time today."

[0066] As shown in Figure 11, the interviewee inputs "Answer 9", which is an answer to "Question 9", to the interviewee terminal 14. The interviewee terminal 14 outputs "Answer 9" to the interview device 12. In the example shown in Figure 11, "Answer 9" is "Thank you very much."

[0067] In this manner, as the exchange of questions by the interview device 12 and answers by the interviewee progresses, a final answer, "Answer N" (in the examples shown in Figs. 3 to 11, N=9), is obtained from the interviewee.

[0068] When the final answer "Answer N" is input to the interview device 12, the input unit 28B generates a prompt, which is a text representing "Questions 1 to N", "Answers 1 to N", and "an instruction to summarize Questions 1 to N and Answers 1 to N", and inputs the generated prompt to the text generation model 58 as an input sentence. The input unit 28B may include an instruction to include an evaluation item for the interviewee in the interview summary. This allows the evaluation content identified from "Answer 1 to N" for the evaluation item to be included in the interview summary. In the example shown in FIG. 11, the input unit 28B generates a prompt saying, "In response to the question "Thank you", the student's answer is "Thank you very much". Please create an interview summary that includes information such as the student's humanity, enthusiasm, and sociability according to the above questions and their answers. " and inputs the generated prompt to the text generation model 58 as an input sentence. The text generated in this case is an example of the "second text" in this disclosure.

[0069] Sentence generation model 58 generates, as an output sentence, an "interview summary" that summarizes "answers 1 to N" and "questions 1 to N" in response to the prompt input by input unit 28B.

[0070] The output unit 28C outputs the "interview summary" as an output sentence generated by the sentence generation model 58. The output unit 28C may output the "interview summary" to the storage 32 and have it recorded in the storage 32, or may output it to another server and have it recorded in the other server. The output unit 28C may also output the "interview summary" to a display and have it displayed on the display, or may output it to a printer and have it printed on the printer.

[0071] The input unit 28B may use an emotion identification model 60 described later and include the emotion of the interviewee determined by the emotion identification model 60 based on the answer of the interviewee in the prompts shown in Figures 3 to 11. This allows the sentence generation model 58 to generate the next question according to the emotion shown by the interviewee in response to the current question.

[0072] Furthermore, the input unit 28B may use an emotion identification model 60 described below and include the emotion of the interviewee determined by the emotion identification model 60 based on the interviewee's answer in the prompt shown in Fig. 6. This allows the emotion shown by the interviewee in response to each question to be included in the interview summary as an index for evaluating the interviewee.

[0073] 3 to 11, a question is output from the interview device 12 to the interviewee via the interviewee terminal 14, and the interviewee inputs an answer to the interview device 12 via the interviewee terminal 14, but a question may be output directly from the interview device 12 to the interviewee, and the interviewee may input an answer directly to the interview device 12. Below, questions will be described as being output to the interviewee, without distinguishing between a case where a question is output to the interviewee via the interviewee terminal 14 and a case where a question is output directly to the interviewee.

[0074] Next, an interview method using the interview device will be described.

[0075] An example of the flow of the interview execution process according to this embodiment is shown in Figure 12. In the interview execution process shown in Figure 12, first, in step ST10, the question initiation unit 28A outputs the first question, "Question 1", to the interviewee.

[0076] Next, in step ST12, the input unit 28B generates a prompt, which is text representing "Question 1," "Answer 1," and "An instruction to ask a question about Answer 1," and inputs the generated prompt as an input sentence to the sentence generation model 58. As a result, the sentence generation model 58 generates the next question, "Question N-1."

[0077] Next, in step ST14, the processor determines whether or not a termination condition for terminating the interview has been met. Examples of termination conditions include a condition that it has been decided that the interview should be terminated based on a policy regarding the progress of the interview, a condition that the number of exchanges of answers and questions has reached a predetermined number, and a condition that the interview time has reached a predetermined time. If the termination condition is met, the interview execution process proceeds to step ST20. On the other hand, if the termination condition is not met, the interview execution process proceeds to step ST16.

[0078] In step ST16, the input unit 28B generates a prompt, which is text representing "Question N-1," "Answer N-1," and "An instruction to ask a question about Answer N-1," and inputs the generated prompt as an input sentence to the sentence generation model 58. As a result, the sentence generation model 58 generates the next question, "Question N."

[0079] Next, in step ST18, the output unit 28C outputs to the interviewee the "question N" generated by the sentence generation model 58. After step ST18, the process returns to step ST14. At this time, N is incremented.

[0080] In step ST20, the input unit 28B generates a prompt, which is text representing "Questions 1 to N", "Answers 1 to N", and "an instruction to summarize Questions 1 to N and Answers 1 to N", and inputs the generated prompt as an input sentence to the sentence generation model 58. As a result, the sentence generation model 58 generates an "interview summary" that summarizes "Answers 1 to N" and "Questions 1 to N".

[0081] Next, in step ST22, the output section 28C outputs "interview summary" as the output sentence generated by the sentence generation model 58. This ends the interview execution process.

[0082] As described above, the interview device 12 makes use of the sentence generation model 58, making it possible to conduct an interview with an interviewee efficiently and accurately.

[0083] Next, the emotion identification model 60 will be described. The emotion identification model 60 may determine the emotion of the user according to a specific mapping. Specifically, the emotion identification model 60 may determine the emotion of the user according to an emotion map (see FIG. 13), which is a specific mapping.

[0084] FIG. 13 is a diagram showing an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive emotions are arranged. The more outside the concentric circles, the more emotions that represent states and actions that arise from a state of mind are arranged. Emotions are a concept that includes emotions and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions that occur in the brain are arranged. On the right side of the concentric circles, emotions that are generally induced by situational judgment are arranged. On the upper and lower sides of the concentric circles, emotions that are generally generated from reactions that occur in the brain and are induced by situational judgment are arranged. In addition, on the upper side of the concentric circles, emotions of "pleasure" are arranged, and on the lower side, emotions of "discomfort" are arranged. In this way, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[0085] These emotions are distributed in the 3 o'clock direction of emotion map 400 and usually fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[0086] The inside of emotion map 400 represents what is going on inside one's mind, and the outside of emotion map 400 represents behavior, so the further out you go on emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[0087] Here, human emotions are based on various balances such as posture and blood sugar level, and when these balances are far from the ideal, it indicates an unpleasant state, and when they are close to the ideal, it indicates a pleasant state. Emotions can also be created for robots, cars, motorcycles, etc., based on various balances such as posture and battery level, so that when these balances are far from the ideal, it indicates an unpleasant state, and when they are close to the ideal, it indicates a pleasant state. The emotion map may be generated, for example, based on the emotion map of Dr. Mitsuyoshi (Research on speech emotion recognition and emotion brain physiological signal analysis system, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). On the left half of the emotion map, emotions belonging to an area called "reaction" where sensation is dominant are lined up. On the right half of the emotion map, emotions belonging to an area called "situation" where situation recognition is dominant are lined up.

[0088] The emotion map defines two emotions that promote learning. The first is the negative emotion around the middle of "repentance" or "remorse" on the situation side. In other words, this is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the positive emotion around "desire" on the response side. In other words, this is when the robot has positive feelings such as "I want more" or "I want to know more."

[0089] The emotion identification model 60 inputs user input to a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the emotion of the user. This neural network is pre-trained based on multiple learning data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 14. FIG. 14 shows an example in which multiple emotions, "relief," "calm," and "encouraging," have similar emotion values.

[0090] In the above embodiment, an example was given in which interview execution processing is performed by a single computer 22, but the technology disclosed herein is not limited to this, and distributed processing of interview execution processing may be performed by multiple computers, including computer 22.

[0091] Furthermore, in the above embodiment, an example has been described in which the interview execution program 56 is stored in the storage 32, but the technology of the present disclosure is not limited to this. For example, the interview execution program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a Universal Serial Bus (USB) memory. The interview execution program 56 stored in the non-transitory storage medium is installed in the computer 22 of the interview device 12. The processor 28 executes the interview execution process in accordance with the interview execution program 56.

[0092] In addition, the interview execution program 56 may be stored in a storage device such as a server connected to the interview device 12 via the network 54, and the interview execution program 56 may be downloaded and installed on the computer 22 upon request from the interview device 12.

[0093] It is not necessary to store the entire interview execution program 56 in a storage device such as a server connected to the interview device 12 via the network 54, or to store the entire interview execution program 56 in the storage 32; only a portion of the interview execution program 56 may be stored.

[0094] The hardware resources for executing the interview execution process can be various processors, as shown below. An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing the interview execution process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, which is a processor having a circuit configuration designed specifically for executing a specific process, such as a Field-Programmable Gate Array (FPGA), a Programmable Logic Device (PLD), or an Application Specific Integrated Circuit (ASIC). Each processor has a built-in or connected memory, and each processor uses the memory to execute the interview execution process.

[0095] The hardware resource for executing the interview execution process may be one of these various processors, or may be a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource for executing the interview execution process may be a single processor.

[0096] As an example of a configuration using a single processor, first, there is a configuration in which one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes the interview execution process. Secondly, there is a configuration in which a processor is used that realizes the functions of the entire system, including multiple hardware resources that execute the interview execution process, on a single IC chip, as typified by SoC (System-on-a-chip). In this way, the interview execution process is realized using one or more of the various processors mentioned above as hardware resources.

[0097] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor elements. The above interview execution process is merely an example. It goes without saying that unnecessary steps may be deleted, new steps may be added, or the order of processing may be changed without departing from the spirit of the invention.

[0098] The above description and illustrations are detailed descriptions of the parts related to the technology of the present disclosure, and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, function, action, and effect is an example of the configuration, function, action, and effect of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above description and illustrations, within the scope of the gist of the technology of the present disclosure. In addition, in order to avoid confusion and to facilitate understanding of the parts related to the technology of the present disclosure, the above description and illustrations omit explanations of technical common sense that do not require explanation in order to enable the implementation of the technology of the present disclosure.

[0099] All publications, patent applications, and standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or standard was specifically and individually indicated to be incorporated by reference.

[0100] The following supplementary notes are disclosed below regarding this embodiment. (Appendix 1) an input unit that inputs a first text representing an answer of an interviewee and an instruction to ask a question about the answer as an input sentence to a sentence generation model that generates an output sentence according to the input sentence; an output unit that outputs a question as the output sentence generated by the sentence generation model to the interviewee; An interview device comprising: (Appendix 2) the input unit inputs, to the sentence generation model, a second text representing the answer, the question, and an instruction to summarize the answer and the question as the input sentence; The output unit outputs an interview summary as the output sentence generated by the sentence generation model. The interview device described in Appendix 1. (Appendix 3) The input unit includes, in the second text, an instruction to include evaluation items for the interviewee in the interview summary. The interview device described in Appendix 2. (Appendix 4) The evaluation items include at least one of the interviewee's personality, enthusiasm, and sociability. The interview device described in Appendix 3. (Appendix 5) The input unit determines a policy regarding the progress of the interview based on the answer. An interview device described in any one of Appendix 1 to Appendix 4. (Appendix 6) The input unit determines a policy regarding the progress of the interview based on the evaluation items for the interviewee. An interview device described in any one of appendix 1 to appendix 5. (Appendix 7) The input unit includes in the first text an emotion of the interviewee determined by an emotion identification model based on the answer. An interview device described in any one of appendix 1 to appendix 6. (Appendix 8) The input unit includes in the second text an emotion of the interviewee determined by an emotion identification model based on the answer. An interview device described in any one of Appendix 2 to Appendix 4 and Appendix 5 to Appendix 7 dependent on Appendix 2. (Appendix 9) The computer inputting a first text representing an answer of an interviewee and an instruction to ask a question in response to the answer as the input sentence into a sentence generation model that generates an output sentence in response to the input sentence; and outputting a question as the output sentence generated by the sentence generation model to the interviewee; An interview method that involves carrying out a process. (Appendix 10) On the computer, inputting a first text representing an answer of an interviewee and an instruction to ask a question in response to the answer as the input sentence into a sentence generation model that generates an output sentence in response to the input sentence; and outputting a question as the output sentence generated by the sentence generation model to the interviewee; A program for executing a process including the above. [Explanation of symbols]

[0101] 10. Interview System 12 Interview Equipment 14 Interviewee terminal 22 Computers 24 Database 26 Communication I / F 28 processors 28A Question start part 28B Input section 28C Output section 30RAM 32 Storage 34 Bus 36 Computer 38 Reception equipment 38A Keyboard 38B Mouse 38C Touchpad 38D Microphone 40 Output Device 40A Display 40B Speaker 42 Camera 44 Communication I / F 46 processors 48 RAM 50 Storage 52 Bus 54 Network 56 Interview Implementation Program 58 Text Generation Model 60 Emotion identification model 400 Emotion Map 900 Emotion Map

Claims

1. an input unit that inputs a first text representing an answer of an interviewee and an instruction to ask a question about the answer as the input sentence to a sentence generation model that generates an output sentence according to the input sentence; an output unit that outputs a question as the output sentence generated by the sentence generation model to the interviewee; An interview device comprising:

2. the input unit inputs, to the sentence generation model, a second text representing the answer, the question, and an instruction to summarize the answer and the question as the input sentence; The output unit outputs an interview summary as the output sentence generated by the sentence generation model. The interview device according to claim 1.

3. The input unit includes, in the second text, an instruction to include evaluation items for the interviewee in the interview summary.

3. An interview device according to claim 2.

4. The evaluation items include at least one of the interviewee's personality, enthusiasm, and sociability.

4. An interview device according to claim 3.

5. The input unit determines a policy regarding the progress of the interview based on the answer. The interview device according to claim 1.

6. The input unit determines a policy regarding the progress of the interview based on the evaluation items for the interviewee. The interview device according to claim 1.

7. The input unit includes in the first text an emotion of the interviewee determined by an emotion identification model based on the answer. The interview device according to claim 1.

8. The input unit includes in the second text an emotion of the interviewee determined by an emotion identification model based on the answer.

3. An interview device according to claim 2.

9. The computer inputting a first text representing an answer of an interviewee and an instruction to ask a question in response to the answer into a sentence generation model that generates an output sentence in response to the input sentence as the input sentence; and outputting a question as the output sentence generated by the sentence generation model to the interviewee; An interview method that involves carrying out a process.

10. On the computer, inputting a first text representing an answer of an interviewee and an instruction to ask a question in response to the answer into a sentence generation model that generates an output sentence in response to the input sentence as the input sentence; and outputting a question as the output sentence generated by the sentence generation model to the interviewee; A program for executing a process including the above.

Images