Acquisition method, generation method, system therefor and program for enabling a dialog between a computer and a human using natural language

Patent metadata

Field	Value
Publication number	US-10964323-B2
Application number	US-201716301627-A
Country	US
Kind code	B2
Filing date	May 19, 2017
Priority date	May 20, 2016
Publication date	Mar 30, 2021
Grant date	Mar 30, 2021

Abstract

An acquisition method is a method performed by an acquisition system in order to acquire a speech set in which three speeches are associated with one another used to generate a second speech made by a dialog system, based on a speech set in which three speeches are associated with one another, in response to a speech made by a human in response to a first speech made by the dialog system. A storage part of the acquisition system stores a plurality of speech sets in which two speeches are associated with each other and the acquisition method includes a presentation step of presenting in order, a speech t(1) and a speech t(2) which are two consecutive speeches included in a certain speech set stored in the storage part of the acquisition system, a speech receiving step of receiving input of a third speech t(3) which is a human speech after presenting the speech t(2) and a storing step of storing the speech t(1), the speech t(2), and the speech t(3) associated with one another as a speech set in which three speeches are associated with one another in the storage part of the acquisition system.

First claim

What is claimed is:

1. An acquisition method executed by an acquisition system to acquire a three-speech set in which three speeches, each of which forms a portion of a dialog that includes a plurality of speakers, are associated with one another used to generate a second speech made by a dialog system in response to a human speech made by a human speaker in response to a first speech made by the dialog system based on a three-speech set recorded in advance in which three speeches are associated with one another, a storage part of the acquisition system storing two-speech sets in advance, each of two-speech sets having two speeches which are associated with each other, the method comprising:
a presentation step of presenting in order, a speech t( 1 ) and a speech t( 2 ) which are two consecutive speeches, each of which forms a portion of a dialog that includes a plurality of speakers, included in a certain two-speech set stored in advance in the storage part of the acquisition system;
a speech receiving step of receiving input of a third speech t( 3 ) which is a human speech made by a human speaker after presenting the speech t( 2 ); and
a storing step of storing the speech t( 1 ) and the speech t( 2 ) recorded in advance and the speech t( 3 ) associated with one another as a three-speech set in which three speeches are associated with one another in the storage part of the acquisition system, wherein
the acquisition system includes a first agent and a second agent different from the first agent, the first agent and the second agent being two different virtual non-human synthesized speakers, and the speech t( 1 ) is presented by the first agent and the speech t( 2 ) is presented by the second agent, and
the speech t( 2 ) is a portion of dialog that responds directly to the speech t( 1 ).

2. The acquisition method according to claim 1 , further comprising a second storing step of storing the speech t( 2 ) associated with the speech t( 3 ) in the storage part of the acquisition system as a two-speech set in which two speeches are associated with each other.

3. An acquisition method executed by an acquisition system to acquire a N-speech set in which N speeches, each of which forms a portion of a dialog that includes a plurality of speakers, are associated with one another where N is a maximum value of the number of speeches associated with a speech and assumed to be any one of an integer equal to or greater than 3, used to generate an N-th speech made by a dialog system in response to an (N−1)-th human speech made by a human speaker after first to (N−2)-th speeches made by the dialog system or/and a human based on a N-speech set recorded in advance in which N speeches are associated with one another, a storage part of the acquisition system storing a (N−1)-speech set in advance including N−1 speeches associated with one another, the acquisition method comprising:
a presentation step of presenting in order, a speech t( 1 ) to a speech t(N−1) which are N−1 consecutive speeches, each of which forms a portion of a dialog that includes a plurality of speakers, included in a certain (N−1)-speech set stored in advance in the storage part of the acquisition system;
a speech receiving step of receiving input of an N-th speech t(N) which is a human speech after presenting the speech t(N−1) which is an (N−1)-th speech; and
a storing step of storing in the storage part of the acquisition system a speech t(N−m p +1) to the speech t(N−1) recorded in advance and the speech t(N) associated with one another for each m p as a m p -speech set in which m p speeches are associated with one another, where P represents the total number of speech sets that are associated with speech t(N), p is an index indicating a speech set that is associated with speech t(N) and represents each integer between 1 and P inclusive and m p is associated with p, is the number of speeches included in the p-th speech set associated with speech t(N) and is an integer equal to or greater than 2 and equal to or less than N, wherein
the acquisition system includes a first agent and a second agent different from the first agent, the first agent and the second agent being two different virtual non-human synthesized speakers, and the speech t(N−2) is presented by the first agent and the speech t(N−1) is presented by the second agent, and
the speech t(N−1) is a speech that responds directly to the speech t(N−2).

4. The acquisition method according to claim 3 , wherein the storing step stores the m p -speech set where at least m p =N.

5. A generation method for a generation system to generate a speech made by a dialog system in response to a human speech made by a human speaker, a storage part of the dialog system storing a speech set in advance in which a first speech presented by an acquisition system, a second speech presented by the acquisition system and a third speech which is a speech of a person a made after presenting the second speech are associated with one another, each speech forming a portion of a dialog that includes a plurality of speakers, the generation method comprising:
a presentation step of presenting a speech t′( 1 ) stored in advance in the storage part of the dialog system;
a speech receiving step of receiving input of a second speech t′( 2 ) which is a speech of a human person b after presenting the speech t′( 1 ); and
a generation step of generating a third speech of a speech set in which a first speech is identical or similar to the speech t′( 1 ) and a second speech is identical or similar to the speech t′( 2 ) of the speech set stored in advance in the storage part of the dialog system as a speech of the dialog system after the speech t′( 2 ), wherein
the acquisition system includes a first agent and a second agent different from the first agent, the first agent and the second agent being two different virtual non-human synthesized speakers, and the first speech is presented by the first agent and the second speech is presented by the second agent, and
the second speech is a speech that responds directly to first speech.

6. A generation method for a generation system to generate a speech made by a dialog system in response to a human speech made by a human, a storage part of the dialog system storing a speech set in advance in which a first speech to an (N−1)-th speech where N is assumed to be any one of an integer equal to or greater than 3, made between a person a and an acquisition system and an N-th speech which is a speech of the person a made after the (N−1)-th speech are associated with one another, each speech forming a portion of a dialog that includes a plurality of speakers, the generation method comprising:
a speech receiving step of receiving input of an m-th speech t′(m) which is a speech of a human person b where m is assumed to be any one of an integer equal to or greater than 2 and less than N; and
a generation step of generating at least a speech t(N−j+1) included in speeches following an m consecutive speech t(N−m+1−j) to speech t(N−j) included in a speech set stored in advance in the storage part of the dialog system when the m consecutive speech t(N−m+1−j) to speech t(N−j) are identical or similar to a first speech t′( 1 ) to an m-th speech t′(m) made between the person b and the dialog system, as a speech of the dialog system after the speech t′(m), where N is the number of speeches included in speech sets stored in the storage part, m is the number of speeches used for searching, and j is a number to identify a starting point for the search and is assumed to be any one of an integer equal to or greater than 1 and equal to or less than N−m, wherein
the acquisition system includes a first agent and a second agent different from the first agent, the first agent and the second agent being two differen

Assignees

• Nippon Telegraph & Telephone
• Univ Osaka

Inventors

• Sugiyama Hiroaki
• Meguro Toyomi
• Yamato Junji
• Yoshikawa Yuichiro
• Ishiguro Hiroshi

Classifications

• G10L15/22 Primary

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

• G10L13/00

  Speech synthesis; Text to speech systems · CPC title

• G10L15/1815

  Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning · CPC title

• G10L2015/223

  Execution procedure of a spoken command · CPC title

• A63H11/00 Primary

  Self-movable toy figures · CPC title