Dialogue management device, dialogue management system, and dialogue management method

Abstract

To provide dialogue management means capable of swiftly identifying parts of interest in the transcript of dialogue even when the topic is developed in various ways.SOLUTION: A dialogue management device (210) provides a dialogue interface for digital dialogue with multiple speakers. The dialogue management device includes: a dialogue interface unit (212) that acquires a dialogue record of the digital dialogue; a topic management department unit (214) receives input of topic identification information for identifying a topic in the digital dialogue during the digital dialogue; and a dialogue editorial unit (216) that generates an edited dialogue record organized for each topic by editing the dialogue record based on the topic identification information.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a dialogue management device, a dialogue management system, and a dialogue management method. [Background technology]

[0002] In recent years, online meetings and interviews have been conducted via communication networks such as the Internet. In particular, due to the spread of teleworking, it has become common to hold meetings and interviews online without meeting face-to-face.

[0003] Various functions have been proposed for such online conference systems (online interview systems). For example, a function is known that records and saves videos of conferences held on the online conference system, allowing users to watch them again after the conference has ended. When reviewing such recorded conference videos, it is desirable to be able to quickly identify parts of interest and to make viewing easier.

[0004] One example of a related technology for online interview systems is JP 2021-175155 A (Patent Document 1). Patent Document 1 discloses a technology in which "a mobile phone device (smartphone 100) having an imaging unit and a communication unit functions as an interview unit. The communication unit is configured to be able to communicate with other devices via a telephone network and the Internet. The interview unit has an interview control unit. When the telephone number of the interview partner is input, the interview control unit transmits connection information for live streaming of the interview to the interview partner via the communication unit and the telephone network. When the interview partner connects via the connection information, the transmission and reception of audio in the live streaming of the interview with the interview partner via the communication unit and the telephone network is carried out using the telephone unit of the mobile phone device. The interview control unit controls the transmission and reception of images in the live streaming of the interview with the interview partner via the communication unit and the Internet." [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent Publication No. 2021-175155 Summary of the Invention [Problem to be solved by the invention]

[0006] Patent Document 1 discloses a means for converting the contents of a call into text based on recorded data of a live broadcast of an interview, and analyzing the resulting text data to determine the atmosphere of the interview based on what the interview participants say, etc. Based on the results of this analysis, the participants can modify their facial expressions, speaking style, speaking content, etc. at the next interview, and conduct the interview in a better atmosphere.

[0007] However, Patent Document 1 does not anticipate organizing records of conversations held over the telephone, online conference systems, etc., by topic. In order to obtain a record of a conversation that progresses by topic, it is conceivable to record each topic as the conversation is being carried out, but in reality, in a conversation held by multiple people, the topic may jump to another topic, or return from another topic to the original topic, and the topic may develop in various ways. For this reason, system When trying to organize conversations using a conversation log, it can be difficult to analyze topics or different topics can be mixed together. As a result, there is a problem in that it is difficult to quickly identify parts of interest when viewing conversation logs.

[0008] Therefore, an object of the present disclosure is to provide a dialogue management means that can quickly identify interesting parts in a dialogue record, even when the topic develops in various ways. [Means for solving the problem]

[0009] In order to solve the above problems, a representative dialogue management device of the present invention includes a dialogue interface unit that provides a dialogue interface for conducting digital dialogue between a plurality of speakers and acquires a dialogue record of the digital dialogue, a topic management unit that accepts input of topic designation information for designating a topic in the digital dialogue during the digital dialogue, and a dialogue editing unit that generates an edited dialogue record organized by topic by editing the dialogue record based on the topic designation information. Effect of the Invention

[0010] According to the present disclosure, it is possible to provide a dialogue management means that is capable of quickly identifying interesting parts in a dialogue record, even when the topic develops in various ways. Other objects, configurations and effects will become apparent from the following description of the preferred embodiment of the invention. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 illustrates a computer system for implementing an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a diagram illustrating an example of a configuration of a dialogue management system according to an embodiment of the present disclosure. [Diagram 3] FIG. 3 is a diagram illustrating an example of the flow of a conversation record generation process for each topic according to an embodiment of the present disclosure. [Figure 4] FIG. 4 is a diagram illustrating an example of a dialogue interface and a dialogue record portion for each topic according to an embodiment of the present disclosure. [Diagram 5] FIG. 5 is a diagram illustrating an example of the flow of an interaction interface drawing process according to an embodiment of the present disclosure. [Figure 6] FIG. 6 is a diagram for explaining the topic shaping process and the related information shaping process according to the embodiment of the present disclosure. [Figure 7] FIG. 7 is a diagram illustrating an example of the flow of the dialogue recording timing correction process according to the embodiment of the present disclosure. [Figure 8]FIG. 8 is a diagram illustrating an example of the flow of a useful topic transition data registration process according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Note that the present invention is not limited to the embodiment. In addition, in the description of the drawings, the same parts are denoted by the same reference numerals.

[0013] As described above, it is desirable to quickly identify a portion of interest in a record of a conversation conducted by telephone, an online conference system, or the like. Therefore, in the present disclosure, a menu (topic selection list) for selecting a current topic from previously prepared candidates during a dialogue is displayed in an interface, and input specifying a topic tag indicating the current topic and a start time and end time of the current topic is accepted from the participants of the dialogue. Based on this input, a dialogue record labeled for each topic can be obtained. Then, by editing this dialogue record labeled for each topic and grouping together parts (dialogue record parts) corresponding to the same topic, an edited dialogue record organized by topic can be generated. This makes it possible to quickly identify parts of the conversation record that are of interest, even when the topic develops in various ways.

[0014] First, referring to Fig. 1, a computer system 100 for implementing the embodiments of the present disclosure will be described. The mechanisms and devices of the various embodiments disclosed herein may be applied to any suitable computing system. The main components of the computer system 100 include one or more processors 102, memory 104, terminal interface 112, storage interface 113, I / O (input / output) device interface 114, and network interface 115. These components may be interconnected via a memory bus 106, an I / O bus 108, a bus interface unit 109, and an I / O bus interface unit 110.

[0015] Computer system 100 may include one or more general purpose programmable central processing units (CPUs) 102A and 102B, collectively referred to as processors 102. In some embodiments, computer system 100 may include multiple processors, and in other embodiments, computer system 100 may be a single CPU system. Each processor 102 executes instructions stored in memory 104 and may include an on-board cache.

[0016] In one embodiment, memory 104 may include random access semiconductor memory, storage devices, or storage media (either volatile or non-volatile) for storing data and programs. Memory 104 may store all or a portion of the programs, modules, and data structures that implement the functions described herein. For example, memory 104 may store a dialogue management application 150. In one embodiment, dialogue management application 150 may include instructions or descriptions that execute on processor 102 the functions described below.

[0017] In some embodiments, dialogue management application 150 may be implemented in hardware via semiconductor devices, chips, logic gates, circuits, circuit cards, and / or other physical hardware devices instead of or in addition to a processor-based system. In some embodiments, dialogue management application 150 may include data other than instructions or descriptions. In some embodiments, cameras, sensors, or other data input devices (not shown) may be provided to communicate directly with bus interface unit 109, processor 102, or other hardware of computer system 100.

[0018] Computer system 100 may include a bus interface unit 109 that provides communication between processor 102, memory 104, display system 124, and I / O bus interface unit 110. I / O bus interface unit 110 may couple to an I / O bus 108 for transferring data to and from various I / O units. I / O bus interface unit 110 may communicate via I / O bus 108 with multiple I / O interface units 112, 113, 114, and 115, also known as I / O processors (IOPs) or I / O adapters (IOAs).

[0019] Display system 124 may include a display controller, a display memory, or both. The display controller may provide video, audio, or both data to display device 126. Computer system 100 may also include one or more sensors or other devices configured to collect data and provide the data to processor 102.

[0020] For example, computer system 100 may include biometric sensors to collect heart rate data, stress level data, etc., environmental sensors to collect humidity data, temperature data, pressure data, etc., and motion sensors to collect acceleration data, movement data, etc. Other types of sensors may also be used. Display system 124 may be connected to a display device 126, such as a standalone display screen, a television, a tablet, or a handheld device.

[0021] The I / O interface unit provides the functionality to communicate with various storage or I / O devices. For example, the terminal interface unit 112 may include a video display, a speaker, 、 User I / O devices 116 may be attached, such as user output devices, such as a television, and user input devices, such as a keyboard, mouse, keypad, touchpad, trackball, buttons, light pen, or other pointing device. A user may use a user interface to enter input data or instructions into user I / O devices 116 and computer system 100, and receive output data from computer system 100, by manipulating the user input devices. The user interface may be displayed on a display device, played through speakers, or printed via a printer, for example, via user I / O devices 116.

[0022] Storage interface 113 may be attached to one or more disk drives or direct access storage device 117 (usually a magnetic disk drive storage device, but may also be an array of disk drives or other storage devices configured to appear as a single disk drive). In one embodiment, storage device 117 may be implemented as any secondary storage device. Contents of memory 104 may be stored in storage device 117 and retrieved from storage device 117 as needed. I / O device interface 114 may provide an interface to other I / O devices such as printers, fax machines, etc. Network interface 115 may provide a communications path to allow computer system 100 and other devices to communicate with each other. This communications path may be, for example, network 130.

[0023] In some embodiments, computer system 100 may be a device that receives requests from other computer systems (clients) without a direct user interface, such as a multi-user mainframe computer system, a single-user system, or a server computer. In other embodiments, computer system 100 may be a desktop computer, a portable computer, a laptop, a tablet computer, a pocket computer, a telephone, a smartphone, or any other suitable electronic device.

[0024] Next, a dialogue management system according to an embodiment of the present disclosure will be described with reference to FIG.

[0025] Fig. 2 is a diagram showing an example of a configuration of a dialogue management system 200 according to an embodiment of the present disclosure. The dialogue management system 200 according to an embodiment of the present disclosure is a system for generating edited dialogue records organized by topic for digital dialogues between multiple people, and as shown in Fig. 2, is mainly composed of a dialogue management device 210, a first dialogue device 220, and a second dialogue device 230. The dialogue management device 210, the first dialogue device 220, and the second dialogue device 230 are connected to each other so as to be able to communicate with each other via a communication network 225 such as the Internet.

[0026] The dialogue management device 210 is a device for providing a platform for digital dialogue between the first dialogue device 220 and the second dialogue device 230, and for generating edited dialogue records of the digital dialogues organized by topic. The digital dialogue here means the sharing of information using words between at least two speakers via a telephone, an online conference system, an app, etc. The digital dialogue here may be, for example, a telephone call, etc., in which only audio information is shared, or a video call, etc., in which both audio information and video information are shared. In addition, the term "speaker" here refers to a person who participates in a digital dialogue. For the sake of convenience, a digital dialogue conducted by two speakers (a first speaker and a second speaker) will be described below as an example, but the present disclosure is not limited thereto, and the number of speakers may be any number.

[0027] As shown in FIG. 2, the dialogue management device 210 includes a dialogue interface unit 212 , a topic management unit 214 , and a dialogue editing unit 216 .

[0028] The dialogue interface unit 212 is a functional unit that provides a dialogue interface for digital dialogue between a first speaker and a second speaker and acquires a dialogue record of the digital dialogue. The dialogue interface here means a GUI (Graphical User Interface) screen that is displayed on a display of the first dialogue device 220 and the second dialogue device 230 described later, for example, and provides information and settings related to the dialogue. As an example, the dialogue interface may include, for example, images of each speaker captured by a camera, a text chat window, camera and microphone setting items, etc. In addition, as described later, the dialogue interface according to the embodiment of the present disclosure may display a menu for selecting a current topic from candidates prepared in advance and related materials related to the current topic.

[0029] The topic management unit 214 is a functional unit that accepts input of topic designation information for designating a topic in a digital dialogue during the digital dialogue. More specifically, the topic management unit 214 may display a topic selection list including a plurality of topic candidates in a dialogue interface generated by the dialogue management device 210, and accept a first topic tag selected from the topic selection list, a first start time of the first topic, and a first end time of the first topic as topic designation information based on an input from a speaker (e.g., an interviewer, etc.).

[0030] Thereafter, the topic management unit 214 may associate a first topic tag with a first dialogue record portion corresponding to the first start time of the accepted first topic and the first end time of the first topic. The topic management unit 214 accepts input of topic designation information from the speaker until the digital dialogue ends, and associates the designated topic tag with the designated dialogue record portion, thereby obtaining a dialogue record labeled for each topic. The dialogue record portion here means a portion of the dialogue record.

[0031] The dialogue editing unit 216 is a functional unit that generates an edited dialogue record organized by topic after the digital dialogue is finished, by editing the dialogue record recorded by the dialogue management device 210 based on the topic designation information acquired by the topic management unit 214. More specifically, the dialogue editing unit 216 may generate an edited dialogue record organized by topic by extracting and combining dialogue record parts associated with the same topic tag (e.g., a first topic tag) from the dialogue record labeled by topic. Here, the expression "extracting and assembling dialogue record portions" means, for example, cutting out dialogue record portions corresponding to each topic tag from the dialogue record using an existing video or audio editing means, and combining the dialogue record portions corresponding to the same topic tag into one group. The dialogue record portions in each group may be arranged in chronological order, or may be arranged based on a predetermined topic priority order.

[0032] Although the overview and main functions of the functional units included in the dialogue management device 210 have been described above, the present disclosure is not limited to this, and the dialogue management device 210 may have other functions as described below. Furthermore, in this embodiment, the term "unit" may include, for example, a combination of hardware resources implemented by a circuit in the broad sense and information processing of software that can be specifically realized by these hardware resources (for example, the dialogue management application 150 shown in FIG. 1). Furthermore, this embodiment handles various information, which is represented by, for example, physical values ​​of signal values ​​representing voltage and current, high and low signal values ​​as a binary bit collection consisting of 0 or 1, or quantum superposition (so-called quantum bits), and communication and calculation can be performed on a circuit in the broad sense.

[0033] The first dialogue device 220 and the second dialogue device 230 are terminals used by speakers participating in a digital dialogue, and connect to a digital dialogue platform provided by the dialogue management device 210 via a communication network 225. The first dialogue device 220 and the second dialogue device 230 are terminals having a function for acquiring and providing audio information and / or video information transmitted and received in a digital dialogue, and may be, for example, a smartphone, a tablet, a personal computer, etc. More specifically, the first dialogue device 220 and the second dialogue device 230 may be configured to display a dialogue interface generated by the dialogue interface unit 212 on a screen and accept information input (for example, input of topic designation information) from a user. For ease of explanation, FIG. 2 shows an example of a configuration of the dialogue management system 200 including two dialogue devices. However, in reality, the number of dialogue devices may be determined according to the number of speakers participating in the digital dialogue.

[0034] According to the dialogue management system 200 configured as described above, by generating edited dialogue records organized by topic, it becomes possible to quickly identify parts of the dialogue record that are of interest, even when the topic is expanded in various ways. In addition, in Figure 2, the dialogue management device 210 has been described as an example in which the functions of the dialogue interface unit 212, the topic management unit 214, and the dialogue editing unit 216 are implemented in the dialogue management device 210 accessible via the communication network 225. However, the present disclosure is not limited to this, and it is also possible to configure the dialogue interface unit 212, the topic management unit 214, and the dialogue editing unit 216 to be implemented by software running on the first dialogue device 220 and the second dialogue device 230.

[0035] Next, a conversation record generation process for each topic according to an embodiment of the present disclosure (hereinafter, may be referred to as a "conversation record generation process for each topic") will be described with reference to FIG.

[0036] 3 is a diagram showing an example of the flow of a topic-specific dialogue record generation process 300 according to an embodiment of the present disclosure. The topic-specific dialogue record generation process 300 shown in FIG. 3 is a process for generating edited dialogue records organized by topic, and is performed by the dialogue interface unit 212, the topic management unit 214, and the dialogue editing unit 216 in the dialogue management device 210 described above.

[0037] First, in step S304, the dialogue interface unit 212 acquires information on the topic selection list 302 and renders the dialogue interface 306 including the acquired topic selection list 302. The topic selection list 302 is information including a plurality of topic candidates. Dialogue The topic candidates included in the topic selection list 302 may be created in advance by the organizer of the digital dialogue, etc. The topic candidates included in the topic selection list 302 may be created, for example, based on agenda information to be discussed in a particular digital dialogue, etc. As an example, when the digital dialogue is a job interview at a company, the topic selection list 302 may include, for example, "interpersonal relationships at a previous job" and "projects handled at a previous job."

[0038] Here, the dialogue interface unit 212 may generate a dialogue interface 306 including the acquired topic selection list 302 in addition to images of each speaker, a text chat window, camera and microphone settings, etc. As described above, this dialogue interface 306 may be displayed on the dialogue devices (the first dialogue device 220 and the second dialogue device 230) of the speakers participating in the digital dialogue.

[0039] Next, in step S312, the dialogue interface unit 212 acquires the dialogue record 308, and the topic management unit 214 acquires the topic designation information 310. Here, a set of information including the dialogue record 308 and the topic designation information 310 is collectively referred to as "dialogue information 314." The dialogue record 308 is information that records a digital dialogue and may include audio information, video information, or both. For example, if the digital dialogue is conducted by telephone, the dialogue record 308 may be a recording of audio information from the speakers who participated in the digital dialogue, and if the digital dialogue is a video call, the dialogue record 308 may be a recording of audio and video information from the speakers who participated in the digital dialogue.

[0040] The topic designation information 310 is information including a first topic tag selected from the topic selection list 302 displayed in the dialogue interface 306, a first start time of the first topic, and a first end time of the first topic.

[0041] Here, the topic management unit 214 may receive, as topic designation information, a first topic tag selected from the topic selection list 302, a first start time of the first topic, and a first end time of the first topic, based on an input from a speaker (e.g., an interviewer) to the topic selection list 302. Thereafter, the topic management unit 214 may associate the first topic tag with a first dialogue record portion corresponding to the received first start time of the first topic and the first end time of the first topic. The topic management unit 214 accepts input of topic designation information from the speaker until the digital dialogue ends, and associates the designated topic tag with the designated dialogue record portion, thereby obtaining a dialogue record labeled for each topic. In this disclosure, a collection of all topic designation information acquired from a particular speaker during a digital dialogue is referred to as a "topic selection history."

[0042] Next, in step S316, the dialogue editing unit 216 generates an edited dialogue record organized by topic based on the dialogue information acquired in step S312. After the digital dialogue is completed, the dialogue editing unit 216 generates an edited dialogue record organized by topic by editing the dialogue record 308 based on the topic designation information 310. More specifically, the dialogue editing unit 216 may generate an edited dialogue record organized by topic by extracting and combining dialogue record portions associated with the same topic tag (e.g., a first topic tag) from the dialogue records labeled by topic.

[0043] According to the topic-specific dialogue record generation process 300 described above, edited dialogue records organized by topic can be generated, so that even when the topic is expanded in various ways, it becomes possible to quickly identify parts of the dialogue record that are of interest.

[0044] Next, a dialogue interface and an edited dialogue record according to an embodiment of the present disclosure will be described with reference to FIG.

[0045] FIG. 4 is a diagram illustrating an example of a dialogue interface 306 and an edited dialogue record 318 according to an embodiment of the present disclosure.

[0046] As shown in Fig. 4, the dialogue interface 306 displays an image 405 of the other person in the digital dialogue and the topic selection list 302 described above. The topic selection list 302 also includes topic tags of topic candidates such as "relationships at a previous job" and "projects handled at a previous job." A speaker (e.g., a first speaker) participating in a digital dialogue using the dialogue interface 306 may select a topic tag (first topic tag) corresponding to a current topic (first topic) while the digital dialogue is progressing in real time. This causes the selected first topic tag to be associated with the current time (first start time of the first topic) in the dialogue record.

[0047] Also, as the digital dialogue progresses, each time the topic changes, the speaker may select a new topic tag (second topic tag) from the topic selection list 302. Then, the selected second topic tag is associated with the current time (the first end time of the first topic and the first start time of the second topic) in the dialogue record, and the first topic is assumed to have ended. Until the digital dialogue is terminated, input of topic designation information is accepted from the speaker, and the designated topic tag is associated with the designated dialogue record portion, thereby obtaining a dialogue record labeled for each topic.

[0048] Furthermore, after the digital dialogue is completed, the dialogue editing unit 216 may extract and compile dialogue record portions that correspond to the same topic tag (e.g., the first topic tag or the second topic tag) from the dialogue record labeled by topic, to generate an edited dialogue record 318 organized by topic as shown in FIG. 4.

[0049] For example, in a dialogue record, a first topic tag "interpersonal relationships at a previous job" may be associated with a first dialogue record portion corresponding to 00:00-02:00 and a third dialogue record portion corresponding to 03:30-04:00, and a second topic tag "projects handled at a previous job" may be associated with a second dialogue record portion corresponding to 02:00-03:30. In this case, the dialogue editing unit 216 may edit the dialogue record to link the first dialogue record portion and the third dialogue record portion associated with the first topic tag in chronological order, link the second dialogue record portion associated with the second topic tag to the rear of the third dialogue record portion, and generate an edited dialogue record 318 organized by topic. The order of the topics in the dialogue record 318 may be, for example, the order in which the topics appeared in the digital dialogue, or may be based on a predetermined priority order.

[0050] Next, with reference to FIG. 5, a flow of an interactive interface drawing process according to an embodiment of the present disclosure will be described.

[0051] 5 is a diagram showing an example of the flow of a dialogue interface drawing process 500 according to an embodiment of the present disclosure. The dialogue interface drawing process 500 shown in FIG. 5 is a process for generating a dialogue interface including the topic selection list 302 and an image of the dialogue partner, and is executed by the dialogue interface unit 212 and the topic management unit 214 in the dialogue management device 210 described above.

[0052] As described above, in one aspect of the present disclosure, during a dialogue, a topic selection list 302 is displayed in a dialogue interface 306 to allow the user to select a current topic from pre-prepared candidates, and topic designation information that specifies a topic tag indicating the current topic and a start time and end time of the current topic is accepted from participants in the dialogue. However, when inputting topic designation information, if a large number of topic candidates are displayed in the topic selection list 302, it may be time-consuming to search for and select a desired topic. Therefore, one aspect of the dialogue interface drawing process 500 according to an embodiment of the present disclosure relates to a topic shaping process for highlighting topic candidates highly relevant to the current topic in the topic selection list 302. This saves the time and effort of searching for and selecting a desired topic, and a participant of the dialogue who inputs topic designation information can easily select a topic candidate highly relevant to the current topic.

[0053] In addition, when a dialogue is conducted, a speaker participating in the dialogue may want to evaluate or analyze a particular content in the dialogue. For example, if the dialogue is a job interview for a company, the interviewer, who is a speaker participating in the dialogue, may want to evaluate or analyze the candidate's Aptitude There may be a case where a person wishes to evaluate a certain evaluation item, such as a person's ability or a certain level of knowledge. Therefore, another aspect of the dialogue interface drawing process 500 relates to displaying related materials (e.g., evaluation criteria for evaluation items) that are highly relevant to the current topic in the dialogue interface 306. This allows the participants in the dialogue to easily check the highly relevant materials while engaging in the dialogue.

[0054] First, in step S510, the topic management unit 214 inputs the topic transition data 502 and the topic selection list 302, and performs a topic shaping process for highlighting topic candidates highly relevant to the current topic in the topic selection list 302 based on the input topic transition data 502 and topic selection list 302. Here, the topic management unit 214 may input previously acquired dialogue information 524 in addition to the above-mentioned input information.

[0055] The topic transition data 502 here is information regarding a transition from one topic (first topic) to another topic (second topic), and may include, for example, a dialogue ID for identifying the current topic, a transition ID for identifying the transition, a topic ID for identifying each topic, dialogue content indicating the content of the dialogue (also referred to as "matching text information" in this disclosure), a topic start time before correction (first start time of the first topic), a topic end time before correction (first end time of the first topic), a topic start time after correction (second start time of the first topic), and a topic end time after correction (second end time of the first topic). The topic selection list 302 has been described above, so a description thereof will be omitted here. The details of the topic shaping process for highlighting topic candidates highly related to the current topic in the topic selection list 302 will be described later, and therefore will not be described here.

[0056] In step S512, the topic management unit 214 inputs the related materials 506 and the evaluation item / topic correspondence table 508, and performs related information formatting processing to display related materials that are highly relevant to the current topic in the dialogue interface 306 based on the input related materials 506 and the evaluation item / topic correspondence table 508. Here, the topic management unit 214 may input previously acquired dialogue information 524 in addition to the above-mentioned input information.

[0057] The related materials 506 here are materials that are created before the digital dialogue and are related to the digital dialogue, and may be, for example, presentation materials related to the agenda of the digital dialogue, reference materials, etc. As an example, if the digital dialogue is a job interview, the related materials 506 may include responses to a questionnaire administered to the candidate, the candidate's resume, etc. The evaluation item-topic correspondence table 508 is information in a table format showing the correspondence between topics and evaluation items. As an example, when the digital dialogue is a job interview, the evaluation item-topic correspondence table 508 may show that the topic "projects handled at a previous job" corresponds to the evaluation item "work experience." The details of the related information shaping process for displaying related materials highly related to the current topic on the dialogue interface 306 will be described later, and therefore will not be described here.

[0058] Next, in step S514, the dialogue interface unit 212 renders the dialogue interface 306 based on the information acquired from the topic shaping process in step S510 and the related information shaping process in step S512. As described above, the dialogue interface unit 212 may generate a dialogue interface 306 including a topic selection list 302 highlighting topic candidates highly relevant to the current topic, and related information highly relevant to the current topic, in addition to images of each speaker, a text chat window, camera and microphone setting items, etc. This dialogue interface 306 may be displayed on the dialogue devices (first dialogue device 220 and / or second dialogue device 230) of the speakers participating in the digital dialogue.

[0059] Next, in step S516, the dialogue interface unit 212 acquires the dialogue record 308 and the participant memo 522, and the topic management unit 214 acquires the topic designation information 310. Here, a set of information including the dialogue record 308, the participant memo 522, and the topic designation information 310 is collectively referred to as “dialogue information” 524. As discussed above, interaction recording 308 is a recorded digital interaction and may include audio information, video information, or both.

[0060] As described above, the topic designation information 310 is information including a first topic tag selected from the topic selection list 302 displayed in the dialogue interface 306, a first start time of the first topic, and a first end time of the first topic. The participant notes 522 are notes entered by speakers participating in the digital dialogue in the dialogue interface 306. The information in the participant notes 522 may be recorded for each topic. The dialogue information 524 acquired here is fed back to the topic shaping process in step S510 and the related information shaping process in step S512. In this way, the topic shaping process and related information shaping process performed thereafter can provide more accurate topic shaping results and related information shaping results by using the dialogue information 524.

[0061] Next, the topic shaping process and the related information shaping process according to the embodiment of the present disclosure will be described with reference to FIG.

[0062] 6 is a diagram for explaining the topic shaping process and the related information shaping process according to an embodiment of the present disclosure. In FIG. 6, an example of a dialogue interface 610 created using the topic shaping process and an example of a dialogue interface 620 created using the related information shaping process are shown.

[0063] As described above, the topic shaping process is a process for highlighting topic candidates that are highly relevant to the current topic in the topic selection list 302, and is performed by the topic management unit 214. Here, the dialogue interface unit 212 acquires a first utterance (e.g., a utterance made within the last 10 seconds) that satisfies a predetermined closeness criterion (e.g., a criterion that specifies a predetermined period going back from the current time), and generates first utterance text information by converting the first utterance into text information. This first utterance may be an utterance made to any of the speakers participating in the digital dialogue. Also, an existing automatic speech recognition means may be used to convert this first utterance into text information.

[0064] Next, the topic management unit 214 determines the degree of association between the first utterance text information and the topic candidates included in the topic selection list 302. Here, as a method for determining the degree of association, one of the following two methods may be used.

[0065] In the first relevance determination method, the topic management unit 214 searches for topic transition data corresponding to the current topic transition from topic transition data collected in advance. After that, the topic management unit 214 compares the first utterance text information with the matching text information included in the searched topic transition data and associated with each of the topic candidates in the topic selection list 302, based on the number of matching words, to calculate the relevance of the first utterance to each of the multiple topic candidates. information is text information that is determined to be highly relevant to a specific topic candidate and is predicted to have a high probability of appearing when discussing that topic, and is prepared in advance; this information is included in the topic transition data described above. In addition, here, in addition to the number of matching words, the relevance between the first utterance text information and the topic candidates included in the topic selection list 302 may be determined using the similarity of semantic information determined by natural language processing.

[0066] In the second relevance determination method, the topic management unit 214 may search for topic transition data corresponding to the current topic transition from topic transition data collected in advance, and output topic candidates corresponding to the next transition destination specified by the searched topic transition data as topic candidates with high relevance. In this way, it is possible to determine the relevance between the content of the current dialogue and the topic candidate, based on the first speech text information and the topic transition data.

[0067] Next, the topic management unit 214 determines topic candidates whose relevance to the first utterance satisfies a predetermined relevance criterion (e.g., the candidate with the highest relevance), and the dialogue interface unit 212 highlights the determined topic candidates in the topic selection list 302.

[0068] 6, in the dialogue interface 610 created using the topic shaping process, the dialogue interface unit 212 may display topic candidates whose relevance to the first utterance meets a predetermined relevance criterion at a top position, may highlight them by changing their color, or may highlight them by adding a text tag such as "recommended", and the highlighting method is not particularly limited. Also, as shown in the dialogue interface 610, the dialogue interface unit 212 may display a text tag such as "candidate" for other topic candidates having a certain relevance.

[0069] As described above, the related information shaping process is a process for displaying related materials that are highly relevant to the current topic in the dialogue interface, and is performed by the topic management unit 214 . Here, similar to the topic shaping process, the topic management unit 214 may acquire a second utterance (e.g., a utterance made within the last 10 seconds) that satisfies a predetermined closeness criterion (e.g., a criterion that specifies a predetermined period going back from the current time) and convert the second utterance into text information to generate second utterance text information. This second utterance may be a utterance made to any of the speakers participating in the digital dialogue. Also, an existing automatic speech recognition means may be used to convert this second utterance into text information.

[0070] Next, the topic management unit 214 calculates the relevance of the second utterance to each of the items included in the related materials by comparing the second utterance text information with related materials related to the current digital dialogue. Here, in order to calculate the relevance of the second utterance to each of the multiple items included in the related materials, the dialogue interface unit 212 may calculate based on the number of words that match the second utterance text information for each item included in the related materials, or may calculate based on the similarity of semantic information obtained by natural language processing.

[0071] Next, the topic management unit 214 determines an item of the related materials whose relevance to the second utterance satisfies a predetermined relevance criterion (e.g., the item with the highest relevance), and the dialogue interface unit 212 displays the determined item in the dialogue interface 306.

[0072] 6, the dialogue interface unit 212 may display, in a dialogue interface 620 created using the related information shaping process, responses to a questionnaire conducted in advance as items 626 of related materials whose relevance to the second utterance satisfies a predetermined relevance criterion. In addition, the dialogue interface unit 212 may display, in the dialogue interface 620, a window for inputting participant notes 522 entered by speakers participating in the digital dialogue.

[0073] Furthermore, in one aspect of the present disclosure, when the information of the related material item 626 to be displayed does not fit within the dialogue interface 620, the dialogue interface unit 212 may automatically scroll the related material item 626 based on the content of the current topic of the digital dialogue. Here, as described above, the dialogue interface unit 212 may convert utterances made within a predetermined time period prior to the current time into text information, and scroll the related material item 626 having the highest number of matching words between the text information and the related material item 626 to a position where it can be easily viewed.

[0074] According to the above-described topic shaping process, when topic designation information is input, even if a large number of topic candidates are displayed in the topic selection list 302, topic candidates highly relevant to the current topic can be highlighted in the topic selection list 302. This saves the trouble of searching for and selecting a desired topic, and enables a participant of the dialogue who inputs topic designation information to easily select a topic candidate highly relevant to the current topic. Furthermore, according to the related information shaping process described above, related materials highly relevant to the current topic can be displayed in the dialogue interface 306. This allows the participants in the dialogue to easily check the highly relevant materials while engaging in the dialogue.

[0075] Next, a dialogue recording timing correction process according to an embodiment of the present disclosure will be described with reference to FIG.

[0076] Fig. 7 is a diagram showing an example of the flow of a dialogue recording timing correction process 700 according to an embodiment of the present disclosure. The dialogue recording timing correction process 700 shown in Fig. 7 is a process for correcting the start and end times (first start and end times) of a topic specified by a speaker who is a participant in a digital dialogue to more accurate times (second start and end times), and is performed by the dialogue editing unit 216 in the dialogue management device 210.

[0077] As described above, in one aspect of the present disclosure, during a dialogue, a topic selection list 302 is displayed in a dialogue interface 306 to allow the user to select a current topic from pre-prepared candidates, and topic designation information that specifies a topic tag indicating the current topic and a start time and end time of the current topic is accepted from participants in the dialogue. However, when entering topic specification information, the start and end times of the topic specified by the speaker may deviate from the actual start and end times of the topic, for example, when the speaker mistakenly selects the start time of a new topic too early, or when the speaker forgets to select the start time of a topic and selects it too late.

[0078] Therefore, one aspect of the dialogue record timing correction process 700 according to an embodiment of the present disclosure relates to correcting the start and end times of a topic specified by a speaker who is a participant of a digital dialogue to a more accurate time. The more accurate time of the change in topic can be determined, for example, by analyzing the speaker's facial expression changes, mouth movements, word changes, etc. shown in the dialogue record. In this way, by correcting the start and end times of the topic to a more accurate time, an edited dialogue record that is more accurately labeled can be generated.

[0079] First, in step S704, the dialogue editing unit 216 calculates a minimum topic time 706 used for topic invalidation judgment, which will be described later, based on the topic transition data 502 and the minimum topic time calculation coefficient 702. When performing the dialogue record timing correction process 700, in order to correct the start and end times of topics specified by speakers who are participants in a digital dialogue to more accurate times, it is desirable to eliminate topics that do not meet a predetermined length threshold (i.e., topics that are too short). The minimum topic time 706 is information indicating a length threshold for distinguishing between valid and invalid topics when performing timing correction. For convenience of explanation, the time from the start time to the end time of a particular topic will be referred to as "topic time" below.

[0080] This minimum topic time 706 may be set in advance by an administrator or the like, or may be calculated from the topic transition data 502 and the minimum topic time calculation coefficient 702. In this case, the dialogue editing unit 216 may determine the shortest topic time from the topic transition data collected in advance based on the pre-correction and post-correction times included in the topic transition data, and calculate the minimum topic time 706 by multiplying the determined shortest topic time by the preset minimum topic time calculation coefficient 702.

[0081] In step S 708 , the dialogue editing unit 216 calculates the pre- and post-correction topic change timing difference 710 based on the topic transition data 502 . When performing the dialogue recording timing correction process 700, it is desirable to correct the start and end times of a topic specified by a speaker who is a participant in a digital dialogue to a topic change timing candidate that has a small difference from the start and end times of the topic specified by the speaker. Therefore, here, the dialogue editing unit 216 calculates the topic change timing difference before and after correction 710 as a threshold for eliminating topic change timing candidates that have a large difference from the start and end times of the topic specified by the speaker from among the topic change timing candidates. Here, the dialogue editing unit 216 may calculate the topic change timing difference before and after correction 710 by averaging the differences between the topic start time and topic end time before correction and the topic start time and topic end time after correction in the past topic transition data 502, and may set the calculated average value as the topic change timing difference before and after correction 710 (time difference threshold).

[0082] In step S714, the dialogue editing unit 216 calculates the topic time for each topic based on the topic selection history 712 of the target dialogue record (i.e., information indicating all topic start times and topic end times specified for the dialogue record). As an example, the dialogue editing unit 216 may determine that the topic time for the first topic is 3 minutes 47 seconds, the topic time for the second topic is 27 seconds, and the topic time for the third topic is 8 minutes 16 seconds, and output this information as the topic time 716.

[0083] In step S718, the dialogue editing unit 216 performs a topic invalidation judgment based on the minimum topic time 706 calculated in step S704 and the topic-by-topic time 716 calculated in step S714. The topic invalidation judgment here is a judgment for distinguishing between valid topics and invalid topics when performing timing correction. The dialogue editing unit 216 compares the topic-by-topic time 716 calculated in step S714 with the minimum topic time 706 calculated in step S704 to judge, for each topic included in the dialogue record, whether or not the topic time of the topic satisfies the minimum topic time 706. Topics that do not satisfy the minimum topic time 706 are excluded as invalid topics, and a valid topic selection history 720 including only topics that satisfy the minimum topic time 706 is output.

[0084] In step S724, the dialogue editing unit 216 analyzes the dialogue record 308 to determine topic change timing candidates 726, which are candidates for new start and end times of the topic (i.e., start and end times that are more accurate than the start and end times of the topic specified by a speaker who is a participant in the digital dialogue). Here, the dialogue editing unit 216 may analyze the dialogue record 308 using a predetermined face recognition means or voice analysis means. For example, the dialogue editing unit 216 may determine the topic change timing candidates 726 by analyzing the speaker's start of speech, changes in facial expressions, and / or changes in words using analysis means such as speaker separation, voice analysis, automatic voice recognition, etc.

[0085] In step S730, the dialogue editing unit 216 judgement A corrected topic selection history 732 is generated based on the pre- and post-correction topic change timing difference 710, the valid topic selection history 720 generated in step S718, and the topic change timing candidate 726 calculated in step S724. More specifically, if there is a topic start or end time in the valid topic selection history 720 whose difference from a topic change timing candidate 726 is within the topic change timing difference before and after correction 710, the dialogue editing unit 216 corrects the topic start or end time to the topic change timing candidate 726.

[0086] As an example, consider a case where the valid topic selection history 720 includes "01:30" as the start time of the first topic specified by the speaker, the topic change timing candidates 726 include "01:28", "02:30", and "03:30" as change timing candidates for the start time of the first topic, and the topic change timing difference before and after correction 710 is "5 seconds." In this case, the difference between the change timing candidate "01:28" and the start time of the first topic specified by the speaker, "01:30", is within the "5 seconds" specified in the topic change timing difference before and after correction 710, so the change timing candidate "01:28" is adopted, and the start time of the first topic from "01:30" is corrected to "01:28". By carrying out the above-described process for all topic start and end times included in the valid topic selection history 720, it is possible to generate a corrected topic selection history 732 in which the start and end times of topics specified by the speakers who are participants in the digital dialogue are corrected to more accurate times.

[0087] Next, in step S316, the dialogue editing unit 216 edits the dialogue record 308 based on the corrected topic selection history 732 generated in step S730, thereby generating an edited dialogue record 318 in which the start and end times of the topics specified by the speakers who are participants in the digital dialogue are corrected to more accurate times and which is organized by topic.

[0088] According to the dialogue recording timing correction process 700 described above, when the start time and end time of the topic specified by the speaker deviates from the actual start time and end time of the topic, the start time and end time of the topic specified by the speaker can be corrected to the times determined by an analysis means such as face recognition or voice analysis, thereby generating an edited dialogue record that is more accurately labeled.

[0089] Next, a useful topic transition data registration process according to an embodiment of the present disclosure will be described with reference to FIG.

[0090] 8 is a diagram showing an example of the flow of a useful topic transition data registration process 800 according to an embodiment of the present disclosure. The useful topic transition data registration process 800 shown in FIG. 8 is a process for registering useful topic transition data, and is performed by the dialogue editing unit 216 in the dialogue management device 210.

[0091] In the dialogue interface drawing process 500 and the dialogue record timing correction process 700 described above, topic transition data 502 indicating information on a transition from one topic (first topic) to another topic (second topic) is used. For example, in the dialogue interface drawing process 500, the topic transition data 502 is used in a topic shaping process for highlighting topic candidates highly related to the current topic in the topic selection list 302. In the dialogue record timing correction process 700, the topic transition data 502 is used to calculate an appropriate topic change timing difference 710 before and after correction. However, if topic transition data 502 with inappropriate topic transitions (e.g., topic transitions that are logically unnatural) is used, the accuracy of the above-mentioned processing may decrease. Therefore, the useful topic transition data registration process 800 according to an embodiment of the present disclosure relates to selecting and registering useful topic transition data that includes appropriate topic transitions and improves the accuracy of the dialogue interface drawing process 500 and the dialogue recording timing correction process 700.

[0092] First, in step S802, the dialogue editing unit 216 generates a dialogue record for each evaluation item 804, which is a dialogue record organized for each evaluation item, based on the edited dialogue record 318, the related materials 506, the evaluation item-topic correspondence table 508, and the participant memo 522. This dialogue record for each evaluation item 804 differs from the edited dialogue record 318 organized for each topic in that it is organized for each evaluation item. More specifically, the dialogue editing unit 216 may identify topics corresponding to the same evaluation item based on the edited dialogue record 318, the related materials 506, the evaluation item-topic correspondence table 508, and the participant memo 522, and may generate a dialogue record for each evaluation item 804 by combining dialogue record portions corresponding to the identified topics. In this dialogue record for each evaluation item 804, when multiple different topics correspond to the same evaluation item, these multiple topics may be combined into one "topic group."

[0093] Next, in step S806, the dialogue editing unit 216 calculates a certainty factor for each evaluation item 808, which is the certainty factor of the topic group for each evaluation item, based on the dialogue record for each evaluation item 804. The certainty factor for each evaluation item 808 here is a measure that indicates, for each topic group in the dialogue record for each evaluation item 804, the usefulness of that topic group for evaluating a predetermined evaluation item.

[0094] Here, as a method for calculating the confidence factor 808 for each evaluation item, one of the following two methods may be used. In the first pass, the dialogue editing unit 216 uses feature amounts (e.g., facial expression features, voice features) from the past dialogue records 804 for each evaluation item and the evaluation results corresponding to the dialogue records 804 for each evaluation item as learning data, XGboost A prediction model is trained using the above. After that, the dialogue record for each evaluation item 804 to be processed is input to the prediction model trained in this way, and the confidence of the prediction model when calculating the feature amount of the dialogue record for each evaluation item 804 (identifying facial expressions and voices for each frame and calculating the number of frames for each facial expression or voice) may be set as confidence for each evaluation item 808.

[0095] In the second way, the dialogue editing unit 216 may use the prediction model learned as described above to calculate new features (e.g., facial expression features, voice features) from the dialogue record for each evaluation item 804, predict the evaluation result, and set the prediction probability of the prediction model as the confidence level for each evaluation item 808.

[0096] Next, in step S810, the dialogue editing unit 216 calculates a certainty factor for each topic 812, which indicates the certainty factor for each topic, based on the certainty factor for each evaluation item 808. As described above, the certainty factor for each evaluation item 808 is a value indicating the certainty factor for each topic group that combines multiple topics corresponding to each evaluation item, but in order to select topic transition data with high usefulness, it is desirable to calculate the certainty factor for each topic. Here, the dialogue editing unit 216 may calculate the confidence level for each topic 812 by averaging the confidence levels for each topic group shown in the confidence level for each evaluation item 808 for each topic.

[0097] As an example, for a first evaluation item, the certainty of a first topic group consisting of a first topic and a third topic is "0.3," and for a second evaluation item, the certainty of a second topic group consisting of a second topic and a third topic is "0.8." In this case, the first topic appears only in the first topic group corresponding to the first evaluation item, so the confidence level is "0.3." The second topic appears only in the second topic group corresponding to the second evaluation item, so the confidence level is "0.8." The third topic appears in both the first topic group corresponding to the first evaluation item and the second topic group corresponding to the second evaluation item, so the confidence level is "0.55," which is the average of "0.3" and "0.8."

[0098] Next, in step S816, the dialogue editing unit 216 selects useful topic transition data 502 based on the topic confidence 812 calculated in step S810, a preset topic confidence transition threshold 814, and the corrected topic selection history 732 generated in the dialogue recording timing correction process 700. Here, the dialogue editing unit 216 generates all combinations using the topics for which the topic confidence 812 was calculated in step S810, and sets the average of the topic confidence 812 for each combination as the topic transition confidence indicating the usefulness of the combination as a topic transition.

[0099] As an example, in the case of a combination of a first topic with a confidence level of "0.3" and a second topic with a confidence level of "0.8", the topic transition confidence level indicating the effectiveness of a transition from the first topic to the second topic is "0.55", which is the average value of "0.3" and "0.8". In the case of a combination of a second topic with a confidence level of "0.8" and a third topic with a confidence level of "0.55", the topic transition confidence level indicating the effectiveness of a transition from the second topic to the third topic is "0.675", which is the average value of "0.8" and "0.55".

[0100] After calculating the topic transition confidence, the dialogue editing unit 216 registers topic combinations having a topic transition confidence equal to or greater than a preset topic-specific confidence transition threshold 814 as valid topic transition data 502, and removes topic combinations having a topic transition confidence less than the preset topic-specific confidence transition threshold 814 from the dialogue recording timing correction process 700. Here, the dialogue editing unit 216 may register the topic transition data 502 by, for example, storing the topic transition data 502 determined to be valid in a database that can be used in the dialogue interface drawing process 500 and the dialogue recording timing correction process 700.

[0101] According to the useful topic transition data registration process 800 described above, useful topic transition data that includes appropriate topic transitions and improves the accuracy of the dialogue interface drawing process 500 and the dialogue recording timing correction process 700 can be selected and registered.

[0102] According to the dialogue management means of the embodiment of the present disclosure, even in the case of a dialogue in which various topics are developed, for example, an edited dialogue record organized by topic can be generated. According to one aspect of the embodiment of the present disclosure, topic candidates highly relevant to the current topic can be highlighted in the topic selection list, which eliminates the need to search for and select a desired topic, and allows a dialogue participant who inputs topic designation information to easily select a topic candidate highly relevant to the current topic. According to one aspect of the embodiment of the present disclosure, related materials (e.g., evaluation criteria for evaluation items, etc.) that are highly relevant to the current topic can be displayed in the dialogue interface, which allows participants in the dialogue to easily check the highly relevant materials while engaging in the dialogue.

[0103] Furthermore, according to one aspect of the embodiment of the present disclosure, the start and end times of a topic designated by a speaker who is a participant in a digital dialogue can be corrected to more accurate times determined, for example, by changes in the speaker's facial expressions, mouth movements, changes in words, etc. shown in the dialogue record, thereby generating an edited dialogue record that is more accurately labeled.

[0104] Although the embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment, and various modifications are possible without departing from the gist of the present invention. [Explanation of symbols]

[0105] 200 Dialogue Management System 210 Dialogue Management Device 212 Dialogue Interface Section 214 Topic Management Department 216 Dialogue Editorial Department 220 First dialogue device 225 Communication Network 230 Second dialogue device 300 Topic-specific dialogue record generation processing 500 Interactive interface drawing process 700 Dialogue recording timing correction processing 800 Useful topic transition data registration processing

Claims

1. A dialogue management device capable of communicating with at least one dialogue device, The dialogue interface section is providing said dialogue device with a dialogue interface for conducting digital dialogue between a plurality of speakers; obtaining a dialogue record of said digital dialogue; The Topic Management Department: receiving, during the digital dialogue, from the dialogue device, input of topic designation information for designating a topic in the digital dialogue based on a topic selection list including a plurality of topic candidates; The Dialogue Editorial Department, generating an edited dialogue record organized by topic by editing the dialogue record based on the topic designation information; generating an evaluation item-by-evaluation item dialogue record organized for each evaluation item based on the edited dialogue record and an evaluation item-topic correspondence table showing the correspondence between the evaluation items and the topics to be evaluated; A dialogue management device comprising:

2. The dialogue record is: including audio or video information or both corresponding to said digital dialogue; 2. A dialogue management device according to claim 1, characterized in that:

3. The topic management unit includes: displaying a topic selection list including a plurality of topic candidates in the dialogue interface; 2. A dialogue management device according to claim 1, characterized in that:

4. The topic management unit, accept, based on an input from the speaker, a first topic tag selected from the topic selection list, a first start time of the first topic, and a first end time of the first topic as the topic designation information; Associating the first topic tag with a first dialogue record portion corresponding to a first start time of the first topic and a first end time of the first topic; 4. A dialogue management device according to claim 3.

5. The dialogue editing unit includes: Analyzing the dialogue record by a predetermined analysis means to determine a second start time of the first topic and a second end time of the first topic; calculating a time difference between the second start time of the first topic and the first start time of the first topic included in the topic designation information by comparing the second start time of the first topic with the first start time of the first topic; Calculating a time difference threshold based on topic transition data indicating information about transitions between topics in past conversations; If the time difference satisfies the time difference threshold, delete the first topic tag from the first dialogue record portion, and associate the first topic tag with a second dialogue record portion corresponding to a second start time of the first topic and a second end time of the first topic.

5. A dialogue management device according to claim 4.

6. The dialogue editing unit includes: extracting dialogue record portions associated with the first topic tag from the dialogue record and compiling the dialogue record portions to generate edited dialogue records organized by topic; 5. A dialogue management device according to claim 4.

7. The analysis means An analysis means configured to analyze any one of a change in facial expression, a movement of the mouth, and a change in words of the speaker.

6. A dialogue management device according to claim 5.

8. A dialogue management system in which a first dialogue device used by a first speaker, a second dialogue device used by a second speaker, and a dialogue management device for conducting a digital dialogue between the first speaker and the second speaker are connected via a communication network, The dialogue management device includes: a dialogue interface unit that provides a dialogue interface for conducting a digital dialogue between the first speaker and the second speaker to the first dialogue device and the second dialogue device, and obtains a dialogue record of the digital dialogue; a topic management unit that receives, during the digital dialogue, an input of topic designation information for designating a topic in the digital dialogue based on a topic selection list including a plurality of topic candidates; a dialogue editing unit that generates an edited dialogue record organized by topic by editing the dialogue record based on the topic designation information, and generates an evaluation item-based dialogue record organized by evaluation item based on the edited dialogue record and an evaluation item / topic correspondence table that indicates the correspondence between the evaluation items to be evaluated and the topics; 13. A dialogue management system comprising:

9. A dialogue management method in a dialogue management device capable of communicating with at least one dialogue device, comprising: providing said dialogue device with a dialogue interface for conducting digital dialogue between a plurality of speakers; displaying a topic selection list in the dialogue interface, the topic selection list including a plurality of topic candidates; initiating a recording of said digital interaction; receiving, based on an input from the speaker, a first topic tag selected from the topic selection list, a first start time of the first topic, and a first end time of the first topic as topic designation information; Associating the first topic tag with a first dialogue record portion corresponding to a first start time of the first topic and a first end time of the first topic; terminating the recording of the digital interaction and obtaining an interaction record; extracting and compiling dialogue record portions associated with the first topic tag from the dialogue record to generate an edited dialogue record organized by topic; 13. A dialogue management method comprising:

Images