# US 12,518,758 B2 · Conversation assistance method using computer

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,518,758 B2
- Document ID: US12518758B2
- Status: Granted
- Kind: B2
- Application number: 18/686,857
- Filed: 2022-03-17
- Published: 2024-11-14
- Granted: 2026-01-06
- Earliest claimed: 2021-08-27
- Assignees: Interactive Solutions Corp.
- Applicants: Interactive Solutions Corp.
- Inventors: Kiyoshi Sekine
- Web page: https://patentgrep.com/patent/US12518758B2

## CPC

- G10L 15/22: Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L 15/08: Speech classification or search
- G10L 15/10: using distance or distortion measures between unknown speech and reference templates
- G10L 15/26: Speech to text systems (G10L15/08 takes precedence)
- G10L 2015/088: {Word spotting}
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

[Problem] Provided is a conversation assistance method using a computer that ensures evaluating a conversation in real time. [Solution] A conversation assistance method using a computer includes a voice analysis step, a word group analysis step, and a remark evaluation step executed by the computer. The voice analysis step is for analyzing a voice related to a conversation and obtaining a voice word that is a word included in the conversation. The word group analysis step is for analyzing which among a plurality of term groups including either one of or both of a word and a phrase the voice word corresponds to. The remark evaluation step evaluates a remark every time a remark related to the conversation is input, based on the term group analyzed in the word group analysis step.

## Claims

### Claim 1

A computer-implemented method for assisting a computer-implemented conversation comprising executing on a processor: a step of
receiving by a microphone, voice data from a conversation;
inputting the voice data into a memory;
analyzing a voice related to a conversation and obtaining a voice word that is a word included in the conversation;
analyzing which among a plurality of term groups the voice word corresponds to, the term group including either one of or both of a word and a phrase;
storing an evaluation point in relation to a relation between a term group of remarks of one side and a term group of remarks of the other side in the memory; and
evaluating a remark every time a remark related to the conversation is input, based on the evaluation point in relation to the relation between the term group of remarks of one side and the term group of remarks of the other side in the memory;
assigning identification information corresponding to a conversation time domain to a term included in the term group; and
determining the conversation time domain to which current conversation time belongs;
wherein a recommended term group that is a term group desired to be used next by using the term group analyzed, the evaluation point and the determined conversation time domain by the computer is displayed on a display.

### Claim 2 (depends on 1)

The method according to claim 1, wherein
the plurality of term groups include any one or two or more term groups of a greeting-related term group, a counterargument-related term group, a description-related term group, a suspicion-related term group, an excuse-related term group, an agreement-related term group, a promise-related term group, and an affirmation-related term group.

### Claim 3 (depends on 1)

The method according to claim 1, further comprising
a step of inputting a scheduled time of the conversation into the computer, wherein
the recommended term group displaying step includes displaying the recommended term group by further using the scheduled time of the conversation.

### Claim 4 (depends on 1)

The method according to claim 1, further comprising
a step of displaying an image related to the term group analyzed in the word group analysis step by the computer.

### Claim 5

A non-transitory computer-readable information recording medium that stores a program, wherein when the program is executed, it performs:
receiving by a microphone, voice data from a conversation;
inputting the voice data into a memory;
analyzing a voice related to a conversation and obtaining a voice word that is a word included in the conversation;
analyzing which among a plurality of term groups the voice word corresponds to, the term group including either one of or both of a word and a phrase;
storing an evaluation point in relation to a relation between a term group of remarks of one side and a term group of remarks of the other side in the memory; and
evaluating a remark every time a remark related to the conversation is input, based on the evaluation point in relation to a relation between the term group of remarks of one side and the term group of remarks of the other side in the memory;
assigning identification information corresponding to a conversation time domain to a term included in the term group; and
determining the conversation time domain to which current conversation time belongs;
wherein a recommended term group that is a term group desired to be used next by using the term group analyzed and the evaluation point and the determined conversation time domain by the computer is displayed on a display.