Apparatus, system, program and method for generating an interview script tailored to the psychology of the interviewee

Abstract

To provide a device which generates an interview script allowing for a motivational interview considering a psychological state of an interviewee.SOLUTION: The interview script generation device includes: means which extracts a keyword from a series of conversation units in an interview script acquired from an interview script database and, on the basis of the keyword, generates a speech step progress tree obtained by associating spoken contents of conversation units in each speech step with respective psychological types of an interviewee along progress of the speech step; means which selects speech of an interviewer which corresponds to acquired psychological type information of the interviewee and is included in a conversation unit to which the extracted keyword belongs, from a speech group of the interviewer searched from the interview scrip database and classified by the psychological types; and means which combines the selected speech of the interviewer as the speech of the interviewer with the generated speech step progress tree to generate an interview scrip for use in a motivational interview.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a technique for generating an interview script. [Background technology]

[0002] In the field of clinical psychology, motivational interviewing has been attracting a great deal of attention. Motivational interviewing is a person-centered counseling approach that awakens motivation and enthusiasm in the interviewee (client) and encourages the interviewee to change their own behavior.

[0003] Currently, research is progressing on implementing communication that can motivate people by utilizing AI (Artificial Intelligence). For example, Non-Patent Document 1 reviews research on technology in which chatbots using AI interact with people and influence their physical activity, eating behavior, etc. The document suggests that reminders, nudges, and feedback provided by chatbots have a significant impact on improving users' physical activity and eating behavior.

[0004] Furthermore, Non-Patent Document 2 reviews a total of 12 papers on empirical research into technologies in which AI-based chatbots affect people's internal states (mental health, well-being, and motivation). In particular, it introduces the results of simple interactions with chatbots, such as the bot spontaneously calling out to the user or nodding when the user is talking, leading to increased motivation to exercise and reduced feelings of anxiety.

[0005] Furthermore, Patent Document 1 discloses a chatbot that communicates with humans, detects a state of affective argumentation during communication, and assists in resolving the conflict. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] US Patent Application Publication No. 2020 / 0410166 [Non-patent literature]

[0007] [Non-Patent Document 1] Jingwen Zhang, Yoo Jung Oh, Patrick Lange, Zhou Yu, & Yoshimi Fukuoka 4, “Artificial intelligence Chatbot behavior change model for designing artificial intelligence Chatbots to promote physical activity and a healthy diet: Viewpoint”. Journal of medical Internet research, 22(9), e22845, 2020 [Non-Patent Document 2] Arielle Aj Scoglio, Erin D Reilly, Jay A Gorman, & Charles E Drebing, “Use of Social Robots in Mental Health and Well-Being Research: Systematic Review”, Journal of medical Internet research, 21(7), e13322, 2019 Summary of the Invention [Problem to be solved by the invention]

[0008] However, even with the use of conventional AI-based technologies such as those described above, there are still significant challenges to conducting appropriate motivational interviews.

[0009] For example, Non-Patent Document 1 introduces a simple (experimental) AI chatbot that assists in motivational interviewing; however, this bot does not take into account the impact of its actions on the inner state of the user (interviewee), and therefore the bot can only make simple utterances and engage in simple non-verbal communication such as nodding.

[0010] Furthermore, the technology disclosed in Non-Patent Document 2, which examines the impact on the inner state, and the technology disclosed in Patent Document 1, which detects the state of emotional disputes, are not yet able to grasp the inner state of the user (interviewee) and provide support and motivation to the user.

[0011] Therefore, an object of the present invention is to provide an interview script generation device, system, program, and method that generates an interview script that enables motivational interviews that also take into account the psychology of the interviewee. [Means for solving the problem]

[0012] According to the present invention, Between the interviewer and the interviewee An interview script generation device for generating an interview script for use in motivational interviewing, The method obtains an existing interview script from an interview script database that stores an existing interview script linked to information on the psychological type of the interviewee, extracts keywords linked to each of a plurality of pre-set speech stages from a series of dialogue units in the existing interview script, and calculates the speech content of the dialogue unit at each of the speech stages for each psychological type of the interviewee based on the extracted keywords. 、 Along with the progression of the speech stage Place them together a speech stage progression tree generating means for generating an associated speech stage progression tree; A group of utterances by an interviewer who has interviewed the interviewee, which are classified by psychological type in the interview script database, Interviewer utterances retrieved from an interview script database from , The psychological type to be classified is Obtained The Psychological information of the interviewee match The extracted keywords Includes On a conversational basis Belongs to an interviewer utterance selection means for selecting an utterance of the person who has been interviewed; Selected interviewee utterances Regarding the psychological category into which the interviewee's utterances were classified Generated speech stage progression tree In the example, the interviewer's utterance is placed in the utterance stage associated with the keyword contained in the dialogue unit to which the interviewer's utterance belongs. , With the interviewer A script generation means for generating an interview script to be used in a motivational interview with the interviewee; Have death, The utterance stage progression tree is generated by selecting the utterance content of the dialogue unit that has the highest occurrence rate of the included keyword in the existing interview script from the utterance content of the dialogue unit that is arranged in one utterance stage and that is linked to the utterance content of the dialogue unit that is arranged in the stage immediately preceding the one utterance stage, and linking the utterance content of the selected dialogue unit with the utterance content of the dialogue unit that is arranged in the utterance stage immediately following the one utterance stage. Characterized by An interview script generator is provided.

[0013] In one embodiment of the interview script generation device according to the present invention, the existing interview scripts stored in the interview script database are interview scripts in which type information relating to the interview strategy used is linked to the utterances of the interviewee, Books interview the script generation device further comprises an interview strategy determination means for determining an interview strategy to be used in the interview from the acquired psychological type information of the interviewee by using a table in which each psychological type is previously associated with one of a plurality of interview strategies; The interviewer utterance selection means is a group of utterances of the interviewee searched from the interview script database, the group of utterances of the interviewee being classified by type of interview strategy for each psychological type, and a keyword extracted from the group of utterances of the interviewee. Includes On a conversational basis Belongs to Based on the psychological information of the interviewee and the interview strategy decided upon, match Select the utterances of the interviewee It is also preferable.

[0014] Furthermore, the interview script generating device according to the present invention comprises: A feature amount is extracted from the existing interview script stored in the interview script database using an encoder of a natural language processing model, and a feature amount used in a past motivational interview is obtained. Features of the interview script and the correct answer Interview Strategy andThe results are used to train a training data set that includes the interview strategy estimation model. , extracted The feature amount is input, and the interview strategy estimation model is Output information on interview strategies and their results from , Based on the output interview strategy and the information on its results, it is assumed that the specified results will be achieved. It is also preferable to further comprise an interview strategy labeling means for determining type information of an interview strategy to be applied and applying the determined type information of the interview strategy to the existing interview script.

[0015] It is also preferred that the multiple interview strategies of the present invention are strategies represented by a combination of one of the motivational strategies of questioning, affirmation, reflection, and summary, and one of the persuasive strategies of authority, commitment, social proof, liking, reciprocity, and scarcity.

[0016] Furthermore, it is also preferable that the psychology type information according to the present invention includes psychology characteristic information relating to which of a plurality of pre-set static psychology types the subject falls under, and psychology state information at the time of the interview relating to which of a plurality of pre-set dynamic psychology types the subject falls under.

[0017] Furthermore, it is also preferred that the speech stage according to the present invention is a change talk stage consisting of a speech stage relating to hopes for change, a speech stage relating to abilities and prospects for change, a speech stage relating to benefits of changing, a speech stage relating to worries and concerns about not changing, and a speech stage relating to specific plans and ideas for change.

[0018] According to the present invention, there is also provided a method for producing a medicament for a medicament comprising the steps of: Between the interviewer and the interviewee An interview script generation system for generating an interview script for use in motivational interviewing, comprising: The method obtains an existing interview script from an interview script database that stores an existing interview script linked to information on the psychological type of the interviewee, extracts keywords linked to each of a plurality of pre-set speech stages from a series of dialogue units in the existing interview script, and calculates the speech content of the dialogue unit at each of the speech stages for each psychological type of the interviewee based on the extracted keywords. 、 Along with the progression of the speech stage Place them togethera speech stage progression tree generating means for generating an associated speech stage progression tree; A group of utterances by an interviewer who has interviewed the interviewee, which are classified by psychological type in the interview script database, Interviewer utterances retrieved from an interview script database from , The psychological type to be classified is Obtained The Psychological information of the interviewee match The extracted keywords Includes On a conversational basis Belongs to an interviewer utterance selection means for selecting an utterance of the person who has been interviewed; Selected interviewee utterances Regarding the psychological category into which the interviewee's utterances were classified Generated speech stage progression tree In the example, the interviewer's utterance is placed in the utterance stage associated with the keyword contained in the dialogue unit to which the interviewer's utterance belongs. , With the interviewer A script generation means for generating an interview script to be used in a motivational interview with the interviewee; Have death, The utterance stage progression tree is generated by selecting the utterance content of the dialogue unit that has the highest occurrence rate of the included keyword in the existing interview script from the utterance content of the dialogue unit that is arranged in one utterance stage and that is linked to the utterance content of the dialogue unit that is arranged in the stage immediately preceding the one utterance stage, and linking the utterance content of the selected dialogue unit with the utterance content of the dialogue unit that is arranged in the utterance stage immediately following the one utterance stage. Characterized by An interview script generation system is provided.

[0019] According to the present invention, further comprising: Between the interviewer and the interviewee An interview script generation program for causing a computer to generate an interview script for use in motivational interviewing, comprising: The method obtains an existing interview script from an interview script database that stores an existing interview script linked to information on the psychological type of the interviewee, extracts keywords linked to each of a plurality of pre-set speech stages from a series of dialogue units in the existing interview script, and calculates the speech content of the dialogue unit at each of the speech stages for each psychological type of the interviewee based on the extracted keywords. 、 Along with the progression of the speech stage Place them together a speech stage progression tree generating means for generating an associated speech stage progression tree; A group of utterances by an interviewer who has interviewed the interviewee, which are classified by psychological type in the interview script database, Interviewer utterances retrieved from an interview script database from , The psychological type to be classified is Obtained The Psychological information of the interviewee match The extracted keywords Includes On a conversational basis Belongs to an interviewer utterance selection means for selecting an utterance of the person who has been interviewed; Selected interviewee utterances Regarding the psychological category into which the interviewee's utterances were classified Generated speech stage progression tree In the example, the interviewer's utterance is placed in the utterance stage associated with the keyword contained in the dialogue unit to which the interviewer's utterance belongs. , With the interviewer A script generation means for generating an interview script to be used in a motivational interview with the interviewee; to make the computer function 、 The utterance stage progression tree is generated by selecting the utterance content of the dialogue unit that has the highest occurrence rate of the included keyword in the existing interview script from the utterance content of the dialogue unit that is arranged in one utterance stage and that is linked to the utterance content of the dialogue unit that is arranged in the stage immediately preceding the one utterance stage, and linking the utterance content of the selected dialogue unit with the utterance content of the dialogue unit that is arranged in the utterance stage immediately following the one utterance stage. Characterized by An interview script generator is provided.

[0020] According to the present invention, furthermore, Between the interviewer and the interviewee 1. A computer-implemented method for generating an interview script for use in motivational interviewing, comprising: The method obtains an existing interview script from an interview script database that stores an existing interview script linked to information on the psychological type of the interviewee, extracts keywords linked to each of a plurality of pre-set speech stages from a series of dialogue units in the existing interview script, and calculates the speech content of the dialogue unit at each of the speech stages for each psychological type of the interviewee based on the extracted keywords. 、 Along with the progression of the speech stage Place them together generating an associated speech stage progression tree; A group of utterances by an interviewer who has interviewed the interviewee, which are classified by psychological type in the interview script database, Interviewer utterances retrieved from an interview script database from , The psychological type to be classified is Obtained The Psychological information of the interviewee match The extracted keywords Includes On a conversational basis Belongs to selecting an utterance from the person interviewed; Selected interviewee utterances Regarding the psychological category into which the interviewee's utterances were classified Generated speech stage progression tree In the example, the interviewer's utterance is placed in the utterance stage associated with the keyword contained in the dialogue unit to which the interviewer's utterance belongs. , With the interviewer generating an interview script to be used in a motivational interview with the interviewee; Have death, The utterance stage progression tree is generated by selecting the utterance content of the dialogue unit that has the highest occurrence rate of the included keyword in the existing interview script from the utterance content of the dialogue unit that is arranged in one utterance stage and that is connected to the utterance content of the dialogue unit that is arranged in the stage immediately preceding the one utterance stage, and connecting the utterance content of the selected dialogue unit with the utterance content of the dialogue unit that is arranged in the utterance stage immediately following the one utterance stage. Characterized by A method for generating an interview script is provided. Effect of the Invention

[0021] According to the interview script generation device, system, program, and method of the present invention, it is possible to generate an interview script that enables a motivational interview that also takes into account the psychology of the interviewee. [Brief description of the drawings]

[0022] [Figure 1] A functional block diagram showing the functional configuration of one embodiment of the interview script generation device according to the present invention. [Diagram 2] FIG. 2 is a schematic diagram showing a specific example of an utterance stage progression tree according to the present invention; [Diagram 3] 1 is a schematic diagram showing one specific example of an interview script according to the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0023] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

[0024] [Interview script generator] FIG. 1 is a functional block diagram showing the functional configuration of an embodiment of an interview script generation device according to the present invention.

[0025] The interview script generation device 1 shown in Figure 1 is one embodiment of an interview script generation device according to the present invention, and uses an interview script obtained from an interview script database (DB) 2, which stores and manages a group of existing interview scripts, and incorporates information related to the psychological characteristics and psychological state of the interviewee to create an "interview script" (data that compiles multiple pairs of interviewer's utterances and interviewee's utterances as the interview progresses) that enables a motivational interview that also takes into account the psychology of the interviewee.

[0026] Here, the generated interview script may be used, for example, for a chatbot that conducts motivational interviews (in which case the "interviewer" may be a chatbot), or may be used as an auxiliary or support means for a human interviewer to conduct motivational interviews with the interviewee, either face-to-face or remotely via the web, etc.

[0027] In any case, in order to generate such an interview script, the interview script generation device 1 specifically: (A) an utterance stage progress tree generating unit 111 for acquiring an existing interview script from an interview script DB2 that stores existing interview scripts linked with the "psychological type information" of the interviewee, extracting keywords linked in advance to each of a plurality of "utterance stages" set in advance from a series of dialogue units in the acquired interview script, and generating a "utterance stage progress tree" that associates the utterance contents of the dialogue units at each "utterance stage" with the progress of the "utterance stage" for each psychological type of the interviewee based on the extracted keywords; (B) an interviewer utterance selection unit 114 for selecting an "interviewer utterance" that corresponds to the acquired "psychological type information" of the interviewee and contains an extracted keyword in the dialogue unit to which it belongs, from a group of "interviewer utterances" searched from the interview script DB2 and classified by psychological type; (C) a script generating unit 115 that generates an "interview script" to be used in a motivational interview between an interviewer and an interviewee by combining the selected "interviewer's utterance" with the generated "utterance stage progress tree" as the interviewer's utterance; The present invention is characterized in that it has the following features.

[0028] Here, as the "utterance stage" of (A) above, various types based on clinical psychological findings can be used. In this embodiment, however, the change talk (DARN-C) stage, i.e., (a) The Desire stage: The stage where the child makes statements such as "I want to be able to do this" (b) Ability stage: Speaking about the ability and prospects for change, such as "I am confident that I can do something" or talking about past success stories. (c) Reason phase: A phase in which participants give reasons why change will have positive consequences. (d) The stage of speech related to worries and concerns about not changing (Need stage): This is the stage where speech is given in which negative reasons are given, such as "If things continue like this, then..."; (e) The Commitment stage, where specific plans and ideas for change are discussed has been adopted.

[0029] By the way, change talk (DARN-C) is a statement that should be elicited from the interviewee (client) in motivational interviewing, and is also called self-motivational statement. In other words, in motivational interviewing, it is important to appropriately elicit the above utterances (change talk) from the interviewee and encourage the interviewee to change his or her own way of thinking and behavior.

[0030] The "speech stage progress tree" (A) above is information that associates speech content with each psychological type in accordance with the progression of the "speech stage." Furthermore, the "interviewer's speech" (B) above, which is an important element for generating an "interview script" in accordance with this tree, is also selected to correspond to the interviewee's "psychological type information." In this way, the interview script generation device 1 can generate an "interview script" that enables motivational interviewing that also takes into account the interviewee's psychological type, based on clinical psychological knowledge.

[0031] [Device functional configuration, interview script generation system, program, and method] Also according to the functional block diagram of Figure 1, the interview script generation device 1 of this embodiment has a communication interface unit 101, an interview script storage unit 102, a psychological characteristic information storage unit 103, a terminal log information storage unit 104, a physiological information storage unit 105, a user interface unit 106, and a processor / memory (a calculation processing system with memory function).

[0032] Here, this processor memory stores one embodiment of the interview script generation program according to the present invention, and has computer functionality, and performs interview script generation processing by executing this interview script generation program. For this reason, the interview script generation device 1 may be a device dedicated to interview script generation processing, but it can also be a cloud server, a non-cloud server, a personal computer (PC), a notebook or tablet computer, or even a mobile terminal such as a smartphone, equipped with the interview script generation program according to the present invention.

[0033] Furthermore, the processor / memory includes a speech stage progress tree generating unit 111 including a keyword extracting unit 111a, a psychological trait / state estimating unit 112, an interview strategy determining unit 113, an interviewer utterance selecting unit 114, an interview script generating unit 115, an interview strategy labeling unit 116, a communication control unit 121, and an input / output control unit 122. The functional components described above can be considered as functions realized by executing an interview script generating program stored in the processor / memory. The process flow shown by connecting the functional components of the interview script generating device 1 in Figure 1 with arrows can also be understood as one embodiment of the interview script generating method according to the present invention.

[0034] Incidentally, at least one of the speech stage progress tree generation unit 111, psychological characteristic / state estimation unit 112, interview strategy determination unit 113, interviewer utterance selection unit 114, interview script generation unit 115 and interview strategy labeling unit 116 may be a functional component of another device, and the interview script generation function may be performed by a plurality of devices (e.g. a plurality of servers) as a whole. In this case, the plurality of devices as a whole constitutes the interview script generation system according to the present invention. Furthermore, among these, the interview strategy labeling unit 116 may be performed by a device other than this interview script generation device / system, and this interview script generation device / system may obtain a group of interview scripts on which the interview strategy labeling process has been performed from this other device.

[0035] Also in the functional block diagram of FIG. 1, the interview script storage unit 102 is a database of interview scripts that stores and manages a group of existing interview scripts retrieved from the interview script DB2. (a) Each saved interview script is linked to the type information of the “psychological characteristics” and “psychological state” of the person being interviewed (the interviewee at that time), and (b) The utterances of the interviewer (the interviewer at the time) in each saved interview script are linked to type information related to the “interview strategy” used (at the time).

[0036] The above (a) "psychological characteristics" and "psychological state" will be explained in detail later. On the other hand, the above (b) "interview strategy" can be applied to a wide variety of situations as long as it is based on clinical psychological findings. In this embodiment, (b1) one of the motivational strategies “Question,” “Affirm,” “Reflect,” and “Summarize”; (b2) one of the persuasion strategies: “authority,” “commitment,” “social proof,” “liking,” “reciprocity,” and “scarcity” The strategy is expressed as a combination of the above.

[0037] Among these, the motivational strategy (b1) above is a strategy for the interviewer to make utterances in motivational interviews. Here, "questioning" refers to making utterances that allow the interviewee to respond in various ways, such as "how do you feel," "how do you think," and "what do you want to do." "Acknowledging" refers to making utterances to confirm what is acceptable or usable in the interviewee's utterance. "Reflecting" refers to making utterances such as asking the interviewee to repeat what he or she says exactly, asking for a suggested thought, asking the interviewee to repeat what he or she says in an extreme way, or asking for an ambiguous meaning. "Summarizing" refers to making utterances that pick up what is acceptable or usable in the interviewee's utterance. Meanwhile, the persuasive strategy (b2) above is Cialdini's persuasive strategy, which is widely used in the field of behavior modification.

[0038] Incidentally, although linking of interview strategy type information (above (b)) to existing interview scripts can of course be done manually, in this embodiment it can be done automatically by the interview strategy labeling unit 116. Specifically, the interview strategy labeling unit 116: (a) Extracting features from an existing interview script stored in the interview script storage unit 102 (or the interview script DB2) using an encoder of a natural language processing model; (b) inputting the feature quantity into an interview strategy estimation model (for example, constructed using a neural network) trained using learning data including interview strategies used in past motivational interviews and their outcomes; (c) From the output of this interview strategy estimation model, the type information of the interview strategy to be assigned is determined, and the determined type information of the interview strategy is assigned to the existing interview script, and then the existing interview script is returned to the interview script storage unit 102 (or the interview script DB2). It is possible.

[0039] Also in the functional block diagram of Fig. 1, the psychological characteristic information storage unit 103 acquires, stores, and manages "psychological characteristic information" of each interviewee from the psychological characteristic information management server 3 via the communication interface unit 101 and the communication control unit 121. Here, the "psychological characteristic information" is information relating to which of a plurality of preset static psychological types the interviewee falls into.

[0040] In addition, such "psychological characteristic information" can be obtained, for example, when a company conducts personality questionnaire surveys, motivation surveys, work engagement surveys, etc. on all employees who are potential interviewees, and obtains information about each employee (interviewee). (a) personality information based on the Big Five theory (e.g. scores on openness, conscientiousness, extraversion, agreeableness and neuroticism); (b) information on the motivational characteristics of self-regulation theory (self-enhancement), and (c) Information on the types of motivational focus in the Regulatory Focus Theory (which states that there are two types of motivational focus for achieving goals: promotion focus, which aims for positive outcomes, and prevention focus, which aims for avoiding negative outcomes). and stores and manages at least one of the “psychological characteristic information” in the psychological characteristic information management server 3.

[0041] For example, regarding Employee A, psychological characteristic information such as "he is quiet and a little nervous, and loses motivation when he gets stressed (high neuroticism: Big Five theory). He is also more likely to be motivated when he is praised than when he is scolded (self-control theory). Furthermore, he often makes decisions to avoid losses rather than to gain rewards (risk-averse decision-making tendency: regulatory focus theory)" is stored and managed.

[0042] Also in the functional block diagram of Figure 1, in this embodiment, the terminal log information storage unit 104 obtains and stores / manages the usage logs of each interviewee's PC, smartphone, etc., for example the usage logs of each employee's work PC (for example, the usage status of specific software, attendance information, etc.) from the terminal log management server 4 (managed by the company, for example) via the communication interface unit 101 and the communication control unit 121.

[0043] In addition, the physiological information storage unit 105 obtains, via the communication interface unit 101 and the communication control unit 121, physiological information of each interviewee, such as heart rate waveform, heart rate, number of steps, oxygen saturation (SpO2), etc., over a specified past period for each employee who wears a wearable device for health management purposes (as measurement results using the device), from the physiological information management server 5 (managed by the company, for example), and stores and manages the information.

[0044] <Means for estimating psychological characteristics and states> Also in the functional block diagram of FIG. 1, the psychological characteristic / state estimation unit 112 estimates and determines psychological type information of the interviewee. Here, in this embodiment, the psychological type information is as follows: (A) "Psychological characteristic information" relating to which of a plurality of pre-set static psychological types the person corresponds to; (a) "Mental state information" at the time of the interview (including a predetermined time period before the interview) relating to which of a plurality of dynamic psychological types set in advance corresponds; The information includes:

[0045] Of these, with regard to the above (a) "psychological characteristic information," at least one of (a) personality information (e.g., Big Five score), (b) information on motivation characteristics, and (c) information related to the type of motivation focus about the interviewee stored in the psychological characteristic information storage unit 103 as described above can be extracted and used.

[0046] Furthermore, the psychological trait / state estimation unit 112 may extract the usage log of the interviewee's PC, smartphone, etc. stored in the terminal log information storage unit 104, and estimate "psychological trait information" from the log. For example, the usage log of the interviewee (from a predetermined period in the past) may be input to a psychological trait estimation model (constructed, for example, by a neural network) that is trained using the usage log linked (labeled) to the correct psychological trait information as learning data, and the "psychological trait information" of the interviewee may be estimated / determined from the output of the psychological trait estimation model.

[0047] On the other hand, in this embodiment, the above (A) “mental state information” (of the interviewee at the time of the interview) is (d) Information related to (dynamic) self-efficacy, a concept of social cognitive theory; (e) information related to self-agency, a concept in cognitive neuropsychology; and (f) Information on the level of stress At least one of these is true.

[0048] In this embodiment, the psychological characteristic / state estimation unit 112 extracts the use log of the interviewee's PC, smartphone, etc. stored in the terminal log information storage unit 104 and the physiological information of the interviewee stored in the physiological information storage unit 105, and estimates "psychological state information" from these. Specifically, the use log and / or physiological information of the interviewee during the interview (including a predetermined time period before the interview) is input to a psychological state estimation model (constructed, for example, by a neural network) trained using the use log and / or physiological information linked (labeled) to the correct psychological state information as learning data, and the "psychological state information" of the interviewee can be estimated / determined from the output of this psychological state estimation model.

[0049] <Means for generating speech stage progression tree> Similarly, in the functional block diagram of FIG. 1, the speech stage progression tree generating unit 111 includes: (a) acquiring a group of existing interview scripts (to which psychological characteristics and psychological state type information are linked) from the interview script storage unit 102; (b) extracting keywords associated with a plurality of pre-defined speech stages, in this embodiment, change talk stages, from a series of dialogue units (e.g., a pair of an interviewer's utterance and an interviewee's response to the interviewer's utterance) in the acquired existing interview script group; (c) Based on the extracted keywords, a "speech stage progression tree" is generated that associates the speech content of each dialogue unit in each change talk stage (utterance stage) with the progression of the change talk stage (utterance stage) for each psychological type of the interviewee (the interviewee at the time).

[0050] In this embodiment, the group of existing interview scripts (a) above is obtained by classifying each psychological characteristic and type of psychological state, and then keyword extraction is performed on the group of existing interview scripts for each psychological type (type of psychological characteristic and type of psychological state). As a result, a "speech stage progress tree" can be generated for each psychological type (type of psychological characteristic and type of psychological state).

[0051] In this embodiment, the extracted keywords in (b) are from a group of keywords that are pre-set as words that the interviewee (client) is likely to utter in response to questions from the interviewer at each of the stages of D (Desire), A (Ability), R (Reason), N (Need), and C (Commitment) (change talk) stages. Specifically, using a known (natural language processing) text mining method, the most frequently occurring words (nouns, verbs, adjectives, adjectival verbs, etc.) in the utterances of each dialogue unit and that are included in a pre-set group of keywords may be extracted and used as the keywords. In this embodiment, the "appropriateness" is calculated for each extracted keyword. This "appropriateness" is the occurrence rate of the keyword in the existing interview scripts from which it is extracted.

[0052] FIG. 2 is a schematic diagram showing a specific example of an utterance stage progression tree according to the present invention.

[0053] In this specific example, Fig. 2(A) shows the speech stage progress tree for the psychology type: "personality: ***, self-efficacy: high". Also, Fig. 2(B) shows the speech stage progress tree for the psychology type: "personality: ***, self-efficacy: low".

[0054] In each of the speech stage progression trees, the "summary of the interviewer's utterance" indicated by the keywords extracted for a series of dialogue units is placed in the corresponding (linked) speech stage (stages D, A, R, N, and C). Furthermore, the "summary of the interviewer's utterance" (keywords related to it) placed in each speech stage is connected by a directed arrow with appropriateness (occurrence rate) p. Here, the "summary of the interviewer's utterance" (keywords related to it) of one speech stage and the "summary of the interviewer's utterance" (keywords related to it) of the next speech stage connected by a directed arrow are generated with the appropriateness (occurrence rate) p attached to the arrow after the "summary of the interviewer's utterance" at the origin of the arrow is generated.

[0055] Specifically, according to the speech stage progression tree in Figure 2(A), for an interviewee with "personality: ***, self-efficacy: high," the motivational interview is most likely to progress from a state of "not being motivated" to Stage D: "I want to improve" → Stage A: "I will improve on my own" → Stage R: "I want to improve productivity" → Stage N: "My productivity will fall" → Stage C: "I will improve immediately." Furthermore, conducting interviews along these lines is an ideal way of smoothly progressing to Stage C: "I will improve immediately."

[0056] On the other hand, according to the speech stage progression tree in Figure 2(B), for an interviewee with "personality: ***, self-efficacy: low," it is most likely that the motivational interview will progress from the same "unmotivated" state, starting from Stage D: "I don't know what to do" → Stage A: "I want help" → Stage R: "I want to be recognized" → Stage N: "My income will decrease" → Stage C: "I will improve immediately." Conducting interviews along these lines is also an ideal way of smoothly progressing to Stage C: "I will improve immediately."

[0057] <Methods for deciding interview strategy> Returning to the functional block diagram of Fig. 1, the interview strategy determination unit 113 uses the "psychology type-interview strategy correspondence table" to determine an interview strategy (in this embodiment, a motivation strategy and a persuasion strategy) to be used in the interview from the acquired psychology type information of the interviewee (in this embodiment, psychology characteristic information and psychology state information). Incidentally, the interview strategy determined in this way is sometimes called a prime interview strategy.

[0058] Here, the "psychology type-interview strategy correspondence table" is a correspondence table in which each psychology type (type of psychological trait and psychological state) is previously associated with at least one of a plurality of interview strategies (motivational strategies and persuasion strategies). This correspondence may be performed based on knowledge of clinical psychology or on-site experience. In addition, the psychology type-interview strategy correspondence table can be created by using information on the results of past motivational interviews, i.e., information on the types of interview strategies that were successful for interviewees of each psychology type (for example, correlation information between psychology types and types of successful interview strategies).

[0059] The interview strategy determination unit 113 uses such a psychology type-interview strategy correspondence table to present, for example (just as an example) an interviewee who is high in extraversion and high in self-efficacy, a motivation strategy of "approval" and a persuasion strategy of "social proof". Note that it is also preferable that the interview strategy (motivation strategy and persuasion strategy) be determined individually for each speech stage (each stage of DARN-C) for one psychology type information (psychological characteristic information and psychological state information) of the interviewee.

[0060] <Method of selecting interviewer's utterances> Also in the functional block diagram of FIG. 1, the interviewer utterance selection unit 114 selects from the "interviewer utterances" group, which is a group of "utterances of the interviewer (the interviewer at the time)" retrieved from the interview script storage unit 102 and which is classified by type of interview strategy (motivation strategy and persuasion strategy) for each psychological type (type of psychological characteristic and psychological state), (a) The keyword extracted by the keyword extraction unit 111a is included in the dialogue unit to which it belongs (this also applies to dialogue units that include the keyword themselves), and (b) corresponding to the psychological type information (psychological trait information and psychological state information) determined for the interviewee and the interview strategy (motivation strategy and persuasion strategy) determined Select the "interviewer's utterance" that satisfies the condition. Note that, if the interview strategy (motivation strategy and persuasion strategy) is determined individually for each speech stage (each stage of DARN-C), it is also preferable to select the corresponding "interviewer's utterance" for each speech stage.

[0061] As described above, each interview script stored in the interview script storage unit 102 is linked to psychological type information (psychological characteristic information and psychological state information) of the person being interviewed (the interviewee at the time), and type information relating to the interview strategy (motivation strategy and persuasion strategy) used (at the time) is linked to the utterances of the interviewer (the interviewer at the time) in each stored interview script. This enables the interviewer utterance selection unit 114 to create a sub-database of "interviewer utterances" by the selection process described above.

[0062] <Method of generating interview script> Similarly, in the functional block diagram of FIG. 1, the interview script generation unit 115 (a) The "utterance of the interviewer" selected by the interviewer utterance selection unit 114 is set as the "utterance of the interviewer", (b) Among the speech stage progress trees generated by the speech stage progress tree generating unit 111, one corresponding to the psychological type information (psychological characteristic information and psychological state information) of the interviewee is selected; (c) The "interviewer's utterance" in (a) above is combined with the selected speech stage progression tree in (b) above to generate an "interview script" (specific or personalized to the interviewee) to be used in a motivational interview between the interviewer and the interviewee.

[0063] Here, the "interviewer's utterance" (utterance of the person interviewed) is determined individually for each speech stage (each stage of DARN-C), and it is also preferable that it is placed in the corresponding speech stage in the speech stage progression tree selected in (c) above.

[0064] Fig. 3 is a schematic diagram showing a specific example of an interview script according to the present invention. In this example, the psychological type information of the interviewee is "personality: ***, self-efficacy: high", and the speech stage progression tree used is the one shown in Fig. 2(A).

[0065] According to Figure 3, in response to the interviewee's statement at stage A that he or she would "improve on their own," "That's great (acknowledgement). Actually, many people have the same problem as you (social proof). Why do you want to improve it? (This can be a template prepared for the R stage.)" This interviewer's utterance is linked to the interviewee's utterance, "I want to improve my productivity." This increases the possibility that the interviewee will say, "I want to improve my productivity" in the next step, R.

[0066] In addition, in response to the interviewee's statement at stage R that he "want to improve productivity," "What do you mean by "improving productivity"? (Ask again) I understand that it is difficult to ask for change (Goodwill), but on the other hand, will there be inconveniences if things stay as they are? (Here you can use a template prepared in advance for the Nth stage.)" This interviewer's utterance is linked to the interviewee's utterance, "I'm going to be less productive" at the next stage N.

[0067] Furthermore, in response to the interviewee's statement at stage N that "productivity will decrease," "So you're concerned about your productivity (acknowledgement). Actually, there are some tips to improve it. Many people say it works (social proof). Would you like to give it a try? (This could be a pre-prepared template for stage C.)" This interviewer's utterance is linked to the interviewee's utterance of "I will improve immediately." This increases the possibility of the interviewee receiving the final C stage utterance of "I will improve immediately."

[0068] As explained above, the interview script (multiple pairs of interviewer's speech and interviewee's speech (summary)) shown in Figure 3 is suitable for motivational interviewing with interviewees who have the above psychological types. Incidentally, such an interview script is not only expected to improve the effectiveness of motivational interviewing (improving the motivation level of the interviewee and improving their physical and mental health, etc.), but it also makes it possible to reduce the burden on the interviewer.

[0069] Returning to the functional block diagram of FIG. 1, the interview script generated by the interview script generation unit 115 may be transmitted to an external information processing device via the communication control unit 121 and the communication interface unit 101 and used therein. For example, it may be used for a chatbot that performs motivational interviews, or it may be used as an auxiliary / supporting means when a human interviewer conducts motivational interviews with an interviewee face-to-face or remotely via the web, etc. In addition, this generated interview script may be sent to the user interface unit 106 via the input / output control unit 122, displayed on the display unit of the user interface unit 106, and used by the user of the device 1.

[0070] As explained in detail above, in the present invention, the speech stage progression tree is information that associates speech contents with each psychological type along the progression of the speech stage, and the "interviewer's speech" selected according to this tree is also selected to correspond to the psychological type information of the interviewee. In this way, according to the present invention, an interview script that enables motivational interviewing that also takes into account the psychological type of the interviewee can be generated based on clinical psychological findings.

[0071] Also, for example, when providing high-quality and excellent education to children all over the world, it is very important to motivate children to study, but such motivation must be done in a way that suits each individual's personality. In this regard, by generating and utilizing interview scripts for children according to the present invention, it is possible for children to be motivated on their own, and as a result, to enjoy high-quality and excellent education. In other words, according to the present invention, it is also possible to contribute to Goal 4 of the Sustainable Development Goals (SDGs) led by the United Nations, "Ensure inclusive and equitable quality education and promote lifelong learning opportunities for all."

[0072] Furthermore, when providing decent work (jobs that are rewarding and humane) to adults around the world, how to motivate adults to work is also a very important issue. In this regard, by generating and utilizing interview scripts for adults according to the present invention, adults can be motivated spontaneously, and as a result, they can obtain decent work that suits them. In other words, according to the present invention, it is also possible to contribute to Goal 8 of the SDGs led by the United Nations, "Promote inclusive and sustainable economic growth, employment and decent work for all."

[0073] Regarding the various embodiments of the present invention described above, various changes, modifications, and omissions within the scope of the technical idea and perspective of the present invention can be easily made by those skilled in the art. The above description is merely an example and is not intended to be restrictive in any way. The present invention is limited only by the scope of the claims and their equivalents. [Explanation of symbols]

[0074] 1. Interview script generator 101 Communication interface section 102 Interview Script Storage Section 103 Psychological characteristics information storage 104 Terminal log information storage unit 105 Physiological information storage 106 User Interface Section 111 Speech stage progression tree generation unit 111a Keyword Extraction Section 112 Psychological characteristics and state estimation unit 113 Interview Strategy Decision Department 114 Interviewer Speech Selection Section 115 Interview script generation unit 116 Interview Strategy Labeling Department 121 Communication control section 122 Input / Output Control Unit

Claims

1. An interview script generation device for generating an interview script to be used in a motivational interview between an interviewer and an interviewee, comprising: a speech stage progress tree generating means for acquiring an existing interview script from an interview script database that stores an existing interview script linked to psychological type information of an interviewee, extracting keywords linked in advance to each of a plurality of speech stages set in advance from a series of dialogue units in the existing interview script, and generating a speech stage progress tree in which the speech contents of the dialogue units at each of the speech stages are arranged along the progress of the speech stage and associated with each other for each psychological type of the interviewee based on the extracted keywords; an interviewer utterance selection means for selecting, from the group of interviewer utterances searched for in the interview script database, an utterance of the interviewer whose utterances correspond to the psychological type information of the interviewee that has been acquired and which belongs to a dialogue unit including the extracted keyword, said utterance being a group of utterances of the interviewer, said group of utterances being classified by psychological type in the interview script database; a script generating means for generating an interview script to be used in a motivational interview between the interviewer and the interviewee by arranging the selected utterance of the interviewer as an utterance of the interviewer in an utterance stage associated with a keyword contained in a dialogue unit to which the utterance of the interviewer belongs in an utterance stage progression tree generated for the psychological type to which the utterance of the interviewer belongs; having The speech stage progression tree is generated by selecting the speech content of the dialogue unit that has the highest occurrence rate of the included keyword in the existing interview script from the speech content of the dialogue unit that is arranged in one speech stage and that is linked to the speech content of the dialogue unit that is arranged in the stage immediately preceding the one speech stage, and linking the speech content of the selected dialogue unit with the speech content of the dialogue unit that is arranged in the speech stage immediately following the one speech stage. An interview script generation device comprising:

2. The existing interview scripts stored in the interview script database are interview scripts in which type information related to the interview strategy used is linked to the utterances of the interviewer, the interview script generation device further comprises an interview strategy determination means for determining an interview strategy to be used in the interview from the acquired psychological type information of the interviewee, using a table in which each psychological type is previously associated with one of a plurality of interview strategies; The interviewer utterance selection means selects an utterance of the interviewer that belongs to a dialogue unit including an extracted keyword from a group of utterances of the interviewer searched from the interview script database and that corresponds to the psychological type information of the interviewee and the determined interview strategy for each psychological type, and that corresponds to the psychological type information of the interviewee and the determined interview strategy.

2. The interview script generating device according to claim 1 .

3. 3. The interview script generation device according to claim 2, further comprising an interview strategy labeling means for extracting features from the existing interview scripts stored in the interview script database using an encoder of a natural language processing model, inputting the extracted features into an interview strategy estimation model trained using learning data including the features of interview scripts used in past motivational interviews and interview strategies as correct answers and their outcomes, causing the interview strategy estimation model to output information relating to the interview strategies and their outcomes, determining interview strategy type information to be assigned as a predetermined outcome based on the output interview strategy and information relating to its outcome, and assigning the determined interview strategy type information to the existing interview script.

4. The interview script generation device according to claim 2 or 3, characterized in that the multiple interview strategies are represented by a combination of one of the motivation strategies of questioning, affirmation, reflection, and summary, and one of the persuasion strategies of authority, commitment, social proof, liking, reciprocity, and scarcity.

5. An interview script generation device as described in any one of claims 1 to 4, characterized in that the psychological type information includes psychological characteristic information relating to which of a plurality of predetermined static psychological types the interviewee corresponds to, and psychological state information at the time of the interview relating to which of a plurality of predetermined dynamic psychological types the interviewee corresponds to.

6. An interview script generation device as described in any one of claims 1 to 5, characterized in that the speech stage is a change talk stage consisting of a stage of speech related to hopes for change, a stage of speech related to abilities and prospects for change, a stage of speech related to the benefits of changing, a stage of speech related to worries and concerns about not changing, and a stage of speech related to specific plans and ideas for change.

7. An interview script generation system for generating an interview script for use in a motivational interview between an interviewer and an interviewee, comprising: a speech stage progress tree generating means for acquiring an existing interview script from an interview script database that stores an existing interview script linked to psychological type information of an interviewee, extracting keywords linked in advance to each of a plurality of speech stages set in advance from a series of dialogue units in the existing interview script, and generating a speech stage progress tree in which the speech contents of the dialogue units at each of the speech stages are arranged along the progress of the speech stage and associated with each other for each psychological type of the interviewee based on the extracted keywords; an interviewer utterance selection means for selecting, from the group of interviewer utterances searched for in the interview script database, an utterance of the interviewer whose utterances correspond to the psychological type information of the interviewee that has been acquired and which belongs to a dialogue unit including the extracted keyword, said utterance being a group of utterances of the interviewer, said group of utterances being classified by psychological type in the interview script database; a script generating means for generating an interview script to be used in a motivational interview between the interviewer and the interviewee by arranging the selected utterance of the interviewer as an utterance of the interviewer in an utterance stage associated with a keyword contained in a dialogue unit to which the utterance of the interviewer belongs in an utterance stage progression tree generated for the psychological type to which the utterance of the interviewer belongs; having The speech stage progression tree is generated by selecting the speech content of the dialogue unit that has the highest occurrence rate of the included keyword in the existing interview script from the speech content of the dialogue unit that is arranged in one speech stage and that is linked to the speech content of the dialogue unit that is arranged in the stage immediately preceding the one speech stage, and linking the speech content of the selected dialogue unit with the speech content of the dialogue unit that is arranged in the speech stage immediately following the one speech stage. An interview script generation system comprising:

8. An interview script generation program for causing a computer to function to generate an interview script for use in a motivational interview between an interviewer and an interviewee, comprising: a speech stage progress tree generating means for acquiring an existing interview script from an interview script database that stores an existing interview script linked to psychological type information of an interviewee, extracting keywords linked in advance to each of a plurality of speech stages set in advance from a series of dialogue units in the existing interview script, and generating a speech stage progress tree in which the speech contents of the dialogue units at each of the speech stages are arranged along the progress of the speech stage and associated with each other for each psychological type of the interviewee based on the extracted keywords; an interviewer utterance selection means for selecting, from the group of interviewer utterances searched for in the interview script database, an utterance of the interviewer whose utterances correspond to the psychological type information of the interviewee that has been acquired and which belongs to a dialogue unit including the extracted keyword, said utterance being a group of utterances of the interviewer, said group of utterances being classified by psychological type in the interview script database; a script generating means for generating an interview script to be used in a motivational interview between the interviewer and the interviewee by arranging the selected utterance of the interviewer as an utterance of the interviewer in an utterance stage associated with a keyword contained in a dialogue unit to which the utterance of the interviewer belongs in an utterance stage progression tree generated for the psychological type to which the utterance of the interviewer belongs; to make the computer function, The speech stage progression tree is generated by selecting the speech content of the dialogue unit that has the highest occurrence rate of the included keyword in the existing interview script from the speech content of the dialogue unit that is arranged in one speech stage and that is linked to the speech content of the dialogue unit that is arranged in the stage immediately preceding the one speech stage, and linking the speech content of the selected dialogue unit with the speech content of the dialogue unit that is arranged in the speech stage immediately following the one speech stage. An interview script generation program comprising:

9. A computer-implemented interview script generation method for generating an interview script for use in a motivational interview between an interviewer and an interviewee, comprising: a step of acquiring an existing interview script from an interview script database that stores an existing interview script linked to psychological type information of an interviewee, extracting keywords linked to each of a plurality of pre-set speech stages from a series of dialogue units in the existing interview script, and generating a speech stage progression tree in which the speech contents of the dialogue units at each of the speech stages are arranged along the progression of the speech stage and associated with each other for each psychological type of the interviewee based on the extracted keywords; a step of selecting, from the group of utterances of the interviewer searched for in the interview script database, an utterance of the interviewer whose utterance group corresponds to the acquired psychological type information of the interviewee and belongs to a dialogue unit including the extracted keyword, said utterance being a group of utterances of the interviewer, said group being classified by psychological type in the interview script database; generating an interview script to be used in a motivational interview between the interviewer and the interviewee by arranging the selected utterance of the interviewer as the utterance of the interviewer in an utterance stage associated with a keyword contained in a dialogue unit to which the utterance of the interviewer belongs in an utterance stage progression tree generated for the psychological type to which the utterance of the interviewer is classified; having The speech stage progression tree is generated by selecting the speech content of the dialogue unit that has the highest occurrence rate of the included keyword in the existing interview script from the speech content of the dialogue unit that is arranged in one speech stage and that is linked to the speech content of the dialogue unit that is arranged in the stage immediately preceding the one speech stage, and linking the speech content of the selected dialogue unit with the speech content of the dialogue unit that is arranged in the speech stage immediately following the one speech stage. A method for generating an interview script.

Images