Information processing system, information processing apparatus, information processing method, and recording medium

Patent metadata

Field	Value
Publication number	US-10984794-B1
Application number	US-201716336820-A
Country	US
Kind code	B1
Filing date	Sep 13, 2017
Priority date	Sep 28, 2016
Publication date	Apr 20, 2021
Grant date	Apr 20, 2021

Abstract

An information processing system according to an embodiment includes a storage, a plurality of conversationers, and a selector. The storage stores a regulation rule indicating a generation rule of a speech in a conversation. Each of the plurality of conversationers makes conversation with a user on the basis of the regulation rule. The selector selects some or all of the plurality of conversationers to cause a conversation to be performed.

First claim

The invention claimed is:

1. An information processing system comprising:
a storage that stores a regulation rule indicating a generation rule of a speech in a conversation;
a plurality of conversationers that each makes conversation with a user on the basis of the regulation rule; and
a selector that selects some or all of the plurality of conversationers to cause a conversation to be performed, wherein
the storage comprises a scene information storage that stores scene information in which a scene is associated with each of the conversationers, the scene being a situation of the speech, a characterization of the conversation, and/or an intention of the speech, and
the selected some or all of the plurality of conversationers makes the conversation with the user according to the scene information.

2. The information processing system according to claim 1 , further comprising:
a scene determiner that determines a scene of a conversation on the basis of an aspect of a conversation start request of one session from the user,
wherein the selector searches for scene-based correspondence information in the scene information storage using the scene determined by the scene determiner, and selects the some or all of the plurality of conversationers.

3. The information processing system according to claim 1 , further comprising:
an evaluator that evaluates the plurality of conversationers for each of the scenes,
wherein the selector selects the some or all of the plurality of conversationers on the basis of scene-based evaluation information including the evaluation of the plurality of conversationers for each of the scenes by the evaluator.

4. The information processing system according to claim 1 , further comprising:
an evaluator that evaluates the plurality of conversationers for each of the users,
wherein the selector selects the some or all of the plurality of conversationers on the basis of user-based evaluation information including the evaluation of the plurality of conversationers for each of the users by the evaluator.

5. The information processing system according to claim 1 ,
wherein the selector selects the some or all of the plurality of conversationers on the basis of a relevance between user attribute information indicating an attribute of the user who performs conversation and conversationer attribute information indicating an attribute of the conversationer.

6. The information processing system according to claim 1 ,
wherein the selector selects the some or all of the plurality of conversationers on the basis of diversity between user attribute information indicating an attribute of the user who performs conversation and conversationer attribute information indicating an attribute of the conversationer.

7. The information processing system according to claim 5 ,
wherein reference to the user attribute information is limited for the some of the plurality of conversationers.

8. The information processing system according to claim 1 , wherein the generation rule includes a result of machine learning such as a parameter of an activation function varying due to machine learning.

9. The information processing system according to claim 1 , wherein the generation rule uniquely associates the speech and a response in reply to the speech.

10. An information processing apparatus comprising:
a plurality of conversationers that each makes conversation with a user on the basis of a regulation rule stored in a storage in advance, the regulation rule indicating a generation rule of a speech in a conversation; and
a selector that selects some or all of the plurality of conversationers to cause a conversation to be performed,
wherein the selected some or all of the plurality of conversationers makes the conversation with the user according to scene information in which a scene is associated with the selected some or all of the plurality of conversationers, the scene being a situation of the speech, a characterization of the conversation, and/or an intention of the speech.

11. An information processing method, comprising:
storing a regulation rule indicating a generation rule of a speech in a conversation;
storing scene information in which a scene is associated with each of a plurality of conversationers that each make conversation with a user on the basis of the regulation rule
selecting some or all of the plurality of conversationers to cause a conversation to be performed, wherein
the selected some or all of the plurality of conversationer makes the conversation with the user according to the scene information, the scene being a situation of the speech, a characterization of the conversation, and/or an intention of the speech.

12. A non-transitory recording medium storing a program for causing a computer to function as:
a plurality of conversationers that each makes conversation with a user on the basis of a regulation rule stored in a storage in advance; and
a selector that selects some or all of the plurality of conversationers to cause a conversation to be performed,
wherein the selected some or all of the plurality of conversationers makes the conversation with the user according to scene information in which a scene is associated with the selected some or all of the plurality of conversationers, the scene being a situation of the speech, a characterization of the conversation, and/or an intention of the speech.

Assignees

• Toshiba Kk
• Toshiba Digital Solutions Corp

Inventors

• Kaneko Yuki
• Tanaka Yasunari
• Shinozaki Masahisa
• Umeki Hideo
• Yoshida Hisako
• Matsui Ai
• Kanaya Kenryo
• Kubota Keiji
• IGOSHI Kazuho
• Iwasaki Motokazu

Classifications

• G06F40/30 Primary

  Semantic analysis · CPC title

• G10L15/1822

  Parsing for meaning understanding · CPC title

• G06F16/22

  Indexing; Data structures therefor; Storage structures · CPC title

• G09B7/02

  of the type wherein the student is expected to construct an answer to the question which is presented or wherein the machine gives an answer to the question presented by a student · CPC title

• G10L15/22 Primary

  Procedures used during a speech recognition process, e.g. man-machine dialogue · CPC title