JP7231894B1 - Conversation support method using computer - Google Patents

Abstract

A conversation support method, apparatus, program, and storage medium capable of evaluating conversation in real time are provided.
A conversation support method using a computer includes a speech analysis step in which a computer analyzes speech related to conversation and obtains spoken words that are words included in the conversation; a word group analysis process for analyzing which of a plurality of term groups containing either or both of the words belongs to; and a statement evaluation process for evaluating the statement based on the term groups analyzed in the word group analysis process. include. In the utterance evaluation process, each time a utterance related to conversation is input to the computer, the utterance is evaluated, and further, evaluation points can be obtained up to the time when the utterance is made in the conversation.
[Selection drawing] Fig. 1

Description

The present invention relates to a conversation support method using a computer.

Japanese Patent No. 6739811 describes a presentation support device for drawing attention to prohibited words. This device is a device that outputs keywords and alerts associated with NG words. In addition, this device sums up the evaluation points of the keywords, related words, and NG words included in the presentation, and obtains the evaluation points of the entire presentation.

Japanese Patent No. 6739811

The device described above is excellent for evaluating presentations. However, the above device evaluates the entire presentation and does not evaluate the conversation in real time. Therefore, an object of the present invention is to provide a conversation support method capable of evaluating conversations in real time.

The first invention of this specification relates to a conversation support method using a computer. This method includes a speech analysis step, a word group analysis step, and an utterance evaluation step. The speech analysis step is a step in which a computer analyzes speech about conversation and obtains speech words that are words included in the conversation. The word group analysis step is a step in which a computer analyzes which of a plurality of term groups a speech word corresponds to. And the term group includes either or both of words and phrases. The utterance evaluation process is a process in which the computer evaluates an utterance each time a utterance relating to conversation is input based on the term group analyzed in the word group analysis process. In this step, every time an utterance related to conversation is input to the computer, the utterance may be evaluated, and an evaluation point may be obtained up to the time when the utterance is made in the conversation.

A plurality of term groups may be, for example, any one of a greeting term group, a rebuttal term group, an explanation term group, a question term group, an excuse term group, an agreement term group, a promise term group, and an affirmation term group. or includes one or more term groups.

The above method may further include a recommended term group display step for the computer to display a recommended term group, which is a term group that is desired to be used next, using the term group analyzed in the word group analysis step. preferable.

The above method further includes the step of entering a scheduled time for the conversation into the computer;
Preferably, the recommended term group display step displays the recommended term group further using the scheduled conversation time.

Preferably, the above method further comprises an image display step for the computer to display an image of the term group analyzed in the word group analysis step.

The next invention in this specification relates to a program or a computer-readable information recording medium storing the program. This program is a program for causing a computer to implement each of the methods described above.

According to the present invention, it is possible to provide a conversation support method capable of evaluating conversations in real time.

FIG. 1 is a flow chart for explaining a conversation support method using a computer. FIG. 2 is a block diagram of a conversation support device for realizing the conversation support method. FIG. 3 is a conceptual diagram showing an example of voice analysis. FIG. 4 is a table in place of a drawing showing examples of term groups. FIG. 5 is a conceptual diagram showing an example of analyzing word groups. FIG. 6 is a conceptual diagram showing an example of utterance evaluation. FIG. 7 is a conceptual diagram showing an example of real-time evaluation of utterances and conversations. FIG. 8 is a conceptual diagram showing an example of a group of recommended terms.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments for carrying out the present invention will be described below with reference to the drawings. The present invention is not limited to the embodiments described below, and includes appropriate modifications within the scope obvious to those skilled in the art from the following embodiments.

A conversation support method using a computer will be described. FIG. 1 is a flow chart for explaining a conversation support method using a computer. This method includes a speech analysis step (S101), a word group analysis step (S102), and an utterance evaluation step (S103). As shown in FIG. 1, the method may further include either or both of a recommended term group display step (S104) and an image display step (S105).

FIG. 2 is a block diagram of a conversation support device for realizing the conversation support method. As shown in FIG. 2, this device 1 includes a speech analysis section 3, a word group analysis section 5, and an utterance evaluation section . As shown in FIG. 2, either or both of a recommended term group display section 9 and an image display section 11 may be further included. The conversation support device is a computer-based system, and basically each part is an element for implementing each process described above. For example, the speech analysis unit 3 is an element for analyzing speech relating to conversation and obtaining speech words that are words included in the conversation. The word group analysis unit 5 is an element for analyzing which one of a plurality of term groups a speech word corresponds to. The utterance evaluation unit 7 is an element for evaluating a utterance each time a utterance relating to conversation is input based on the analyzed term group. The recommended term group display unit 9 is an element for displaying a recommended term group, which is a term group that is desirable to be used next, using the term group analyzed in the word group analysis step. The image display section 11 is an element for displaying images related to term groups.

A computer has an input section, an output section, a control section, a calculation section, and a storage section, and each element is connected by a bus or the like so that information can be exchanged. For example, the storage unit may store a control program or various types of information. When predetermined information is input from the input unit, the control unit reads out the control program stored in the storage unit. Then, the control unit appropriately reads the information stored in the storage unit and notifies it to the calculation unit. In addition, the control unit conveys appropriately input information to the calculation unit. The calculation unit performs calculation processing using the received various information, and stores the information in the storage unit. The control unit reads the calculation result stored in the storage unit and outputs it from the output unit. In this manner, various processes and steps are executed. Each part and each means execute these various processes.

The device of the present invention may include a terminal connected to a network such as the Internet or an intranet, and a server connected to the network. Of course, a single computer or mobile terminal may function as the device of the present invention, or multiple servers may exist.

Voice analysis step (S101)
The speech analysis step is a step in which a computer analyzes speech about conversation and obtains speech words that are words included in the conversation. Conversations can be mutual conversations, explanations or presentations by speakers to the audience, salespeople and customers, questions or inquiries over the phone, and so on. It may also be a conversation with a telephone attendant. Speech is usually the human voice. Of course, voice includes computer-generated pseudo-voice.
FIG. 3 is a conceptual diagram showing an example of voice analysis. In the example shown in FIG. 3, the conversation between the MR and the doctor (Dr. A) is input to the device. This device may be a device connected to the MR's terminal and the doctor's terminal, or may be one of them. Conversational speech may be input to each terminal through the microphone of the MR's terminal and the microphone of the doctor's terminal, and the conversational speech may be input to the device through each terminal. Also, the microphone of the device may directly input the voice of the MR and the doctor (Dr. A) into the device. The control unit of the apparatus appropriately reads the voice from the storage unit and causes the calculation unit to analyze the input voice. In this way, the device obtains the words (spoken words) involved in the conversation. Such speech analysis devices are known. The device may store the obtained phonetic words in a storage unit.

An example of parsing the spoken words shown in FIG. 3 is as follows.
MR: Sensei, I am always indebted to you.
Today, I asked you to introduce the newly launched DPP4 inhibitor Ipro Tablets.
Doctor: (But) there are a lot of them, but are they on sale again?
I don't plan on using any new DPP4 inhibitors anymore.
MR: Doctor, it is true that several DPP4 inhibitors have already been marketed.
Our company's ipro tablets have advantages over other drugs.
Which DPP4 inhibitor are you currently taking?
Doctor: There are many, so we use them according to the patient's lifestyle.
I wonder if I've been putting out petit tablets a lot lately. Putin has a strong effect just once.

Word group analysis step (S102)
The word group analysis step is a step in which a computer analyzes which of a plurality of term groups a speech word corresponds to. And the term group includes either or both of words and phrases. The storage unit of the device stores the words and phrases included in each term group. Then, the control unit of the apparatus causes the calculation unit to perform calculations for collating the spoken words with the words and phrases included in each term group. In this way, it is analyzed which term group the phonetic word is included in. The analyzed analysis result may be stored in the storage unit.

FIG. 4 is a table in place of a drawing showing examples of term groups. A term group may be a combination of words or phrases grouped into various elements. Also, the term group may be one or more of terms that must be spoken, keywords, and NG terms (terms that must not be spoken). A plurality of term groups may be divided into term groups in which the presence of a certain term group is favorable and term groups in which the existence of a term group is undesirable. In addition, the plurality of term groups are, for example, a term group related to greetings, a term group related to counterarguments, a term group related to explanations, a term group related to doubts, a term group related to excuses, a term group related to consent, a term group related to promises, and a term group related to affirmations. includes any one or more groups of terms of Note that these term groups do not have to be stored in the computer as "greeting term groups" or the like.

FIG. 5 is a conceptual diagram showing an example of analyzing word groups. As shown in FIG. 5, for example, the phrase "I am indebted to you" is included in the term group "greetings." For this reason, when the MR utters, "Teacher, I am always indebted to you.", this utterance is analyzed as a "greeting." Then, regarding this conversation, it is stored in the storage unit that a "greeting" was made.
Next, the doctor says, "(But) there are many, but is it on the market again? I'm not going to use the new DPP4 inhibitor anymore." Then, the combination of the words ``but'' and ``not intended'' included in the doctor's utterance is analyzed as a ``rebuttal''. Then, regarding this conversation, it is stored in the storage unit that the "counterargument" was made after the above-described "greeting".

Speech evaluation step (S103)
The utterance evaluation process is a process in which the computer evaluates an utterance each time a utterance relating to conversation is input based on the term group analyzed in the word group analysis process. This step may be a step of evaluating the utterance each time a utterance related to the conversation is input to the computer and further obtaining evaluation points up to the time when the utterance is made in the conversation. For example, in the storage unit, evaluation points are assigned to each term group. Then, the control unit of the apparatus reads out the evaluation points stored in association with the term group from the storage unit, and evaluates the statement. Note that the device may also evaluate the entire conversation, which is a set of utterances, at the same time. In addition, evaluation points may be stored in the storage unit in association with the relationship between the term group of one statement and the term group of the other statement.

For example, the storage unit stores 0 points for the term group “greeting”. Also, in the storage unit, -1 point is stored for the term group of one utterance: "greeting" and the term group of the other utterance: "rebuttal". The storage unit may store +1 point for the term group of one statement: "counterargument" followed by the term group of the other statement: "counterargument". This device may analyze the terms contained in the dialogue to obtain the coefficients. In this case, for example, when the conversation after the first "counter-argument" contained a term (DDP4 inhibitor) with a factor of 3, the term group in the other statement: 3 × (+1) for the "counter-argument" You may process so that 3 points may be added by . In this case, the device stores the point-adding terms and the point-adding coefficients, and temporarily stores the point-adding coefficients when the conversational text is analyzed and it is found that the point-adding terms are included. , the evaluation score may be multiplied by this point addition coefficient when evaluating a statement. For example, these point-adding terms and point-adding coefficients may be stored in the storage unit in association with materials (for example, presentation materials), and when a conversation about the materials takes place, the point-adding terms and point-adding coefficients related to the materials may be read as appropriate. good.

FIG. 6 is a conceptual diagram showing an example of utterance evaluation.
“MR: Sensei, I am always indebted to you.
Today, I asked you to introduce the newly launched DPP4 inhibitor Ipro Tablets. , the term group: ``Greetings'' is stored in the device. Therefore, the control unit of the apparatus reads out from the storage unit 0 points, which are evaluation points associated with the term group: "greetings". Then, this statement is evaluated as 0 points.

Next, the doctor
“Doctor: (But) there are so many, are they on sale again?
I don't plan on using any new DPP4 inhibitors anymore. ' says.
In relation to this statement, the device stores a group of terms: "counterargument". Then, the control unit of the apparatus reads out from the storage unit -1 point, which is the evaluation point associated with the term group of one statement: "greeting" and the term group of the other statement: "counterargument". Thus, utterances are evaluated in real time.

The device also stores evaluation values so far included in the conversation leading up to this utterance. Then, this device performs an operation to add the evaluation value in this utterance. In this way, the device can obtain evaluation points up to the time when a certain utterance is made in the conversation. Then, the obtained evaluation score of the conversation up to the time when a certain statement is made may be appropriately stored in the storage unit and output.

FIG. 7 is a conceptual diagram showing an example of real-time evaluation of utterances and conversations. As shown in FIG. 7, in this example, for each utterance, the utterance and the conversation leading up to the utterance are evaluated. The content will be displayed in real time.

The above method may further include a recommended term group display step for the computer to display a recommended term group, which is a term group that is desired to be used next, using the term group analyzed in the word group analysis step. preferable. The device selects a term group that should be spoken next after a certain term group, and if a second term group comes after the first term group, a third term that is preferably said next. Memorize groups. Then, the device analyzes the utterance, and uses the information about the utterance, the previous utterance, and the term group to which the previous utterance belongs, to read from the storage unit the term group that is desirable for the next utterance. Then, the device outputs terms included in the read term group that should be said next. Then, the term is displayed on the output unit (for example, monitor) of the speaker's terminal. In this way, the device can assist the speaker to converse.

FIG. 8 is a conceptual diagram showing an example of a group of recommended terms. For example, when a term group A1 is used next to a certain term group, term group B1, term group B2, and term group B3 (or terms selected from these term groups or terms terms included in the group) may be displayed at the speaker's output, or term group B1 and term group C1 may be displayed as the best recommended term groups.

The above method may further include entering into the computer a scheduled time for the conversation. In the recommended term group display step, it is preferable to display the recommended term group further using the scheduled conversation time. The scheduled time of the conversation may be the beginning time of the conversation and the end time of the conversation, or the duration of the conversation (for example, within one hour). Preferably, the device has a clock section for keeping time or time. Then, the storage unit divides the conversation into conversation time areas such as initial, middle, and final. Then, the storage unit assigns identification information according to the conversation time region to the terms included in the term group and stores them. Then, when reading the terms included in the term group, the device can read and output the terms corresponding to the conversation time domain of the conversation.
For example, when it is a predetermined time before the scheduled time of the conversation or when the conversation comes to an end, the conversation end words, which are words for ending the conversation, are stored in the storage unit as a group of recommended words. The ending term can be read and displayed accordingly. In this way, lectures, presentations, etc. can be completed within a predetermined time. Examples of end-of-conversation terms are "to sum up" and "to briefly explain the rest".

Preferably, the above method further comprises an image display step for the computer to display an image of the term group analyzed in the word group analysis step. An example of an image for a term group is a score or an arrow symbol (eg, ↑↓) representing a score for the term group. The storage unit stores images related to these term groups. Then, the device grasps a group of terms related to expression. Then, the device uses the grasped information about the term group to read the image about the term group from the storage unit. The device then outputs the read image. For example, if the image is related to the evaluation value, it is possible to instantly grasp whether the person's remark was positive or negative.

The next invention in this specification relates to a program or a computer-readable information recording medium storing the program. This program is a program for causing a computer to implement each of the methods described above. This program, for example, causes a computer to analyze speech related to conversation, obtain spoken words that are words contained in the conversation, analyze which of a plurality of term groups the spoken word corresponds to, and analyze the analyzed terms. Based on the group, each time an utterance related to conversation is input, it functions to evaluate the utterance.

This invention can be used in the information industry.

1 conversation support device 3 voice analysis unit 5 word group analysis unit 7 utterance evaluation unit 9 recommended term group display unit 11 image display unit

Claims (8)

A conversation support method using a computer,
the computer,
a speech analysis step for analyzing speech relating to a conversation to obtain speech words that are words included in the conversation;
a word group analysis step for analyzing which of a plurality of term groups the spoken word corresponds to, wherein the term group includes either or both words and phrases;
a statement evaluation step of evaluating the statement based on the term group analyzed in the word group analysis step;
The utterance evaluation step is a step of evaluating the utterance each time a utterance related to the conversation is input to the computer and obtaining an evaluation score up to the time when the utterance is made in the conversation.
Method. The method of claim 1, wherein
The plurality of term groups are any of a term group related to greetings, a term group related to counterarguments, a term group related to explanations, a term group related to doubts, a term group related to excuses, a term group related to agreement, a term group related to promises, and a term group related to affirmations. containing one or more term groups,
Method. The method of claim 1, wherein
the computer,
Further comprising a recommended term group display step for displaying a recommended term group, which is a term group that is desirable to be used next, using the term group analyzed in the word group analysis step,
Method. 4. The method of claim 3, wherein
further comprising inputting a scheduled time for the conversation into the computer;
The recommended term group display step further uses the scheduled time of the conversation to display the recommended term group.
Method. The method of claim 1, wherein
the computer,
further comprising an image display step for displaying an image related to the term group analyzed in the word group analysis step;
Method. A conversation support device 1,
a speech analysis unit 3 for analyzing speech relating to conversation and obtaining speech words that are words contained in the conversation;
a word group analysis unit 5 for analyzing which of a plurality of term groups the spoken word corresponds to, wherein the term group includes either or both of words and phrases;
a statement evaluation unit 7 for evaluating the statement based on the term group analyzed by the word group analysis unit 5,
The utterance evaluation unit 7 evaluates the utterance each time a utterance related to the conversation is input to the conversation support device 1, and obtains an evaluation score up to the time when the utterance is made in the conversation.
Device. to the computer,
a speech analysis step for analyzing speech relating to a conversation to obtain speech words that are words included in the conversation;
a word group analysis step for analyzing which of a plurality of term groups the spoken word corresponds to, wherein the term group includes either or both words and phrases;
an utterance evaluation step of evaluating the utterance based on the term group analyzed in the word group analysis step;
The utterance evaluation step is a step of evaluating the utterance each time a utterance related to the conversation is input to the computer and obtaining an evaluation score up to the time when the utterance is made in the conversation.
A program for implementing a method. A computer-readable information recording medium storing the program according to claim 7 .

Images