Apparatus and method for dialogue generation in response to received text

• US 8,856,010 B2
• Filed: 09/17/2009
• Issued: 10/07/2014
• Est. Priority Date: 09/22/2008
• Status: Expired due to Fees

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Doi, Miwako, Kobayashi, Yuka, Yamamoto, Daisuke
• Primary Examiner(s): He, Jialong
• Application Number:
  US12/561,624
• Publication Number:
  US 20100076753A1
• Time in Patent Office: 1,846 Days
• Field of Search: 704/270, 704/275
• US Class Current: 704/275
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A dialog generation apparatus includes a reception unit configured to receive a first text from a dialog partner, an information storage unit configured to store profile information specific to a person who can be the dialog partner and a fixed-pattern text associated with the person, a presentation unit configured to present the first text to a user, a speech recognition unit configured to perform speech recognition on speech the user has uttered about the first text presented to the user, and generate a speech recognition result showing the content of the speech, a generation unit configured to generate a second text from the profile information about the dialog partner, fixed-pattern text about the dialog partner, and the speech recognition result, and a transmission unit configured to transmit the second text to the dialog partner.

Claims

1. A dialogue generation apparatus comprising:

a reception unit configured to receive a first text from a dialogue partner;

an information storage unit configured to store profile information specific to a person who is the dialogue partner and a fixed-pattern text associated with the person;

a presentation unit configured to present the first text to a user;

a speech recognition unit configured to perform speech recognition on speech the user has uttered about the first text presented to the user, and generate a speech recognition result showing content of the speech;

a generation unit configured to generate a second text from the profile information about the dialogue partner, fixed-pattern text about the dialogue partner, and the speech recognition result;

a transmission unit configured to transmit the second text to the dialogue partner;

a communication unit configured to communicate with a household information appliance the user uses, and acquire use history information representing the user'"'"'s use of the household information appliance;

a use history storage unit configured to store the use history information; and

a template storage unit configured to store an additional template including the use history information and an additional text,

wherein the generation unit generates the second text by utilizing the additional text and the result of the speech recognition performed on the user'"'"'s speech about the additional text,

wherein the additional text recites a user'"'"'s way of living, is estimated from the use history information, and is, at least in part, randomly generated.

2. The apparatus according to claim 1, further comprising an input unit configured to receive, from the user, a switching instruction for switching of a method of generating the second text;

wherein the speech recognition unit includes a continuous speech recognition unit configured to perform continuous speech recognition on the speech to acquire a continuous speech recognition result, and an isolated word recognition unit configured to perform isolated word recognition on the speech to acquire an isolated word recognition result, and outputs either the continuous speech recognition result or the isolated word recognition result, as the speech recognition result, in accordance with the switching instruction.

3. The apparatus according to claim 2, further comprising a speech storage unit configured to store the speech as a speech file,

wherein the generation unit generates the second text in association with the speech file, so that the second text is referred to in accordance with the switching instruction.

4. The apparatus according to claim 1, wherein the template storage unit further stores the use history information and a subject name template including the use history information and a subject name associated with the use history information, and the generation unit generates the second text by utilizing the subject name associated with the use history information.

5. The apparatus according to claim 1, wherein the use history information includes at least one of data items representing the use time, type of the appliance the user has used and attribute.

6. A dialogue generation method comprising:

receiving a first text from a dialogue partner;

storing profile information specific to a person who is the dialogue partner and a fixed-pattern text associated with the person;

presenting the first text to a user;

performing speech recognition on speech the user has uttered about the first text presented to the user, and generating a speech recognition result showing content of the speech;

generating a second text from the profile information about the dialogue partner, fixed-pattern text about the dialogue partner, and the speech recognition result;

transmitting the second text to the dialogue partner;

communicating with a household information appliance the user uses, to acquire use history information representing the user'"'"'s use of the household information appliance;

storing the use history information; and

storing an additional template including the use history information and an additional text,

wherein the second text is generated by utilizing the additional text and the result of the speech recognition performed on the user'"'"'s speech about the additional text,

wherein the additional text recites a user'"'"'s way of living, is estimated from the use history information, and is, at least in part, randomly generated.

7. A dialogue generation apparatus comprising:

a communication unit configured to communicate with a household information appliance a user uses, and acquire use history information representing the user'"'"'s use of the household information appliance;

a use history storage unit configured to store the use history information;

a template storage unit configured to store an additional template including the use history information and an additional text;

a speech recognition unit configured to perform speech recognition on speech the user has uttered about the additional text, and generate a speech recognition result showing content of the speech;

a generation unit configured to generate a text for transmission by utilizing the additional text and the speech recognition result; and

a transmission unit configured to transmit the text for transmission to a dialogue partner,

wherein the additional text recites a user'"'"'s way of living, is estimated from the use history information, and is, at least in part, randomly generated.

8. A dialogue generation method comprising:

communicating with a household information appliance a user uses, to acquire use history information representing the user'"'"'s use of the household information appliance;

storing the use history information;

storing an additional template including the use history information and an additional text;

performing speech recognition on speech the user has uttered about the additional text, and generating a speech recognition result showing content of the speech;

generating a text for transmission by utilizing the additional text and the speech recognition result; and

transmitting the text for transmission to a dialogue partner,

wherein the additional text recites a user'"'"'s way of living, is estimated from the use history information, and is, at least in part, randomly generated.