TOP > 外国特許検索 > Voice interaction apparatus, its processing method, and program

Voice interaction apparatus, its processing method, and program

外国特許コード F200010124
整理番号 5788
掲載日 2020年5月18日
出願国 中華人民共和国
出願番号 201810175617
公報番号 108630203
出願日 平成30年3月2日(2018.3.2)
公報発行日 平成30年10月9日(2018.10.9)
優先権データ
  • 特願2017-040580 (2017.3.3) JP
発明の名称 (英語) Voice interaction apparatus, its processing method, and program
発明の概要(英語) The invention provides a voice interaction apparatus, its processing method, and a program. The voice interaction apparatus incudes voice recognition means for recognizing a voice of a user, response-sentence generation means for generating a response sentence to the voice of the user based on the recognized voice, filler generation means for generating a filler word to be inserted in a conversation, output means for outputting the generated response sentence and the generated filler word, and classification means for classifying the generated response sentence into one of predetermined speechpatterns indicating predefined speech types. When the output means outputs, after the user utters a voice subsequent to the first response sentence, the filler word and outputs a second response sentence, the classification means classifies the first response sentence into one of the speech patterns, and the filler generation means generates the filler word based on the speech pattern into whichthe first response sentence has been classified.
従来技術、競合技術の概要(英語) Background Art
There is known a voice interaction device, which is inserted into the filler words (i.e. words of silence in a conversation for filling) to prevent being unnaturally prolonged silence in a conversation (see Japanese Unexamined Patent Application Publication No. 2014-191030 number).
However, the present invention human discovered the following problems. In other words, when the waiting time occurs in the dialog, the speech interactive device on the output form (i.e. of the perfunctory) for filling the silence filler words as a word. Thus, insertion of fill words may not be well suited for the content of conversations (e.g. meaning), thereby causing a dialog becomes unnatural.
特許請求の範囲(英語) [claim1]
1. A voice interaction device, comprising:
Speech recognition means for recognizing the user's voice;
Answer sentence generation means, based on the user's speech by the speech recognition to generate speech recognition apparatus of the reply sentence;
Fill generation means, for generating the user's conversation to be inserted into a fill word in the; and
An output device, configured to output a reply sentence and the sentence generated by the response generation means for generating means generates the fill filler words, wherein,
A voice interaction device further includes a sorting device, by the reply sentence classification device for classifying so as to indicate a pre-defined sentence generated by the response generation means of the speech utterance with one of a predetermined type, and to
When the user should be issued after the 1st output means outputs a language sentence and outputs a 2nd voice following statements in response to fill word,
A language utterance is classified as a 1st sentence classification device according to one of the modes of the application, and to
1st means based on the classification by the fill generation device according to a language of the speech pattern to be classified as a fill word to generate a sentence.

[claim2]
2. The speech interaction apparatus according to claim 1, wherein, the speech interaction apparatus further comprises:
A storage device, for storing table information, table information associated with the utterance including an utterance mode and the type of the feature value with respect to the information of the pattern; and
The feature value calculation means, based on a language-processed by the sorting means with respect to the 1st sentence should be classified as a speech mode of the associated information of the type of the feature value to calculate a characteristic value of a previous or subsequent utterance, wherein,
Based on the feature-value calculating means calculates a fill generation unit generating a fill word to the feature value.

[claim3]
3. The speech interaction apparatus according to claim 2, wherein, with respect to the feature value information of the type comprising at least one of the following: utterance prosody information previously, language information of the speech of the previous, subsequent utterance of the speech of language information and the prosody of the subsequent information.

[claim4]
4. The speech interaction apparatus according to claim 2 or 3, wherein,
A storage device associated with each of the characteristic value stored in the form of a fill style information is filled, each of the filler is filled at least one fill word and indicative of the type including a type of word, and to
1st to by the classification device based on the fill generation device should be classified as a language sentence so as to reduce the number of types of the utterance mode is filled, is selected from the reduced number of the feature-value calculation means calculates a fill style associated with the characteristic values of the one type of fill, a fill type and by selecting the selected word to generate a fill word included in the fill.

[claim5]
5. The method of treatment for a voice interaction device, a voice interaction device comprises:
Speech recognition means for recognizing the user's voice;
Answer sentence generation means, based on the user's speech by the speech recognition to generate speech recognition apparatus of the reply sentence;
Fill generation means, for generating the user's conversation to be inserted into a fill word in the; and
An output device, configured to output a reply sentence and the sentence generated by the response generation means for generating means generates the filler words is filled;
Processing method comprising:
When the user should be issued after the 1st output means outputs a language sentence and outputs a 2nd voice following statements in response to fill word,
A language sentence of the words of the 1st type should be classified as one of a predetermined utterance indicating a predefined mode, and to
A language based on the 1st sentence has been classified as a speech pattern to generate a fill word.

[claim6]
6. A program for a voice interaction device, a voice interaction device comprises:
Speech recognition means for recognizing the user's voice;
Answer sentence generation means, based on the user's speech by the speech recognition to generate speech recognition apparatus of the reply sentence;
Fill generation means, for generating the user's conversation to be inserted into a fill word in the; and
An output device, configured to output a reply sentence and the sentence generated by the response generation means for generating means generates the filler words is filled,
A program for making a computer execute :
When the user should be issued after the 1st output means outputs a language sentence and outputs a 2nd voice following statements in response to fill word,
A language sentence of the words of the 1st type should be classified as one of a predetermined utterance indicating a predefined mode, and to
A language based on the 1st sentence has been classified as a speech pattern to generate a fill word.
  • 出願人(英語)
  • KYOTO UNIVERSITY
  • TOYOTA MOTOR
  • 発明者(英語)
  • NAKANISHI RYOSUKE
  • WATANABE NARIMASA
  • KAWAHARA TATSUYA
  • TAKANASHI KATSUYA
国際特許分類(IPC)
ライセンスをご希望の方、特許の内容に興味を持たれた方は、下記までご連絡ください。

PAGE TOP

close
close
close
close
close
close