Method and apparatus for adding new vocabulary to interactive translation and dialogue systems

Patent metadata

Field	Value
Publication number	US-9576570-B2
Application number	US-84782610-A
Country	US
Kind code	B2
Filing date	Jul 30, 2010
Priority date	Jul 30, 2010
Publication date	Feb 21, 2017
Grant date	Feb 21, 2017

Abstract

The present invention relates to a method and apparatus for adding new vocabulary to interactive translation and dialog systems. In one embodiment, a method for adding a new word to a vocabulary of an interactive dialog includes receiving an input signal that includes at least one word not currently in the vocabulary, inserting the word into a dynamic component of a search graph associated with the vocabulary, and compiling the dynamic component independently of a permanent component of the search graph to produce a new sub-grammar, where the permanent component comprises a plurality of words that are permanently part of the search graph.

First claim

What is claimed is:

1. A method for adding a new word to a vocabulary of an interactive computing system comprising one or more computing devices, the method comprising, with the interactive computing system:
receiving a first input signal, the first input signal comprising speech;
generating a recognition result comprising text corresponding to at least a portion of the input signal;
with the recognition result, identifying a new word to add to the vocabulary of the interactive computing system, the vocabulary comprising a compiled component and an uninstantiated placeholder class in a search graph;
instantiating the placeholder class with the new word;
compiling the instantiated placeholder class independently of the compiled component of the vocabulary;
receiving a second input signal, the second input signal comprising speech; and
comparing the instantiated placeholder class to a portion of the second received input signal.

2. The method of claim 1 , further comprising identifying a pronunciation for the new word; and associating the pronunciation with the new word prior to the compiling.

3. The method of claim 2 , further comprising receiving the pronunciation from a user.

4. The method of claim 2 , further comprising generating a plurality of potential pronunciations based on a spelling of the new word; and selecting one of the plurality of potential pronunciations that best matches the new word as pronounced by a user.

5. The method of claim 1 , further comprising:
associating a translation with the new word.

6. The method of claim 5 , wherein the associating further comprises: receiving the translation from a user.

7. The method of claim 5 , wherein the associating further comprises: retrieving the translation from a dictionary.

8. The method of claim 1 , wherein compiling the instantiated placeholder class produces a new sub-grammar, and the method comprises outputting the new sub-grammar to a recognition processor.

9. The method of claim 1 , wherein compiling the instantiated placeholder class produces a dynamic component of a search graph, and wherein the dynamic component and the compiled component are different components of the search graph.

10. The method of claim 1 , wherein the compiled component comprises a main grammar of the vocabulary, compiling the instantiated placeholder class produces a sub-grammar of the vocabulary, and the sub-grammar is smaller in size than the main grammar.

11. The method of claim 1 , wherein compiling the instantiated placeholder class produces a dynamic component of the vocabulary, and the method comprises rescoring the recognition result using the dynamic component of the vocabulary.

12. A language processing module for adding a new word to a vocabulary, the language processing module embodied in one or more non-transitory computer readable storage media and configured to cause a computing device to:
receive a first input signal, the first input signal comprising speech;
generate a recognition result comprising text corresponding to at least a portion of the input signal;
with the recognition result, identify a new word to add to the vocabulary of the interactive computing system, the vocabulary comprising a compiled component and an uninstantiated placeholder class in a search graph;
instantiate the placeholder class with the new word;
compile the instantiated placeholder class independently of the compiled component of the vocabulary;
receive a second input signal, the second input signal comprising speech; and
compare the instantiated placeholder class to a portion of the second received input signal.

13. The language processing module of claim 12 , further comprising: identifying a pronunciation for the new word; and associating the pronunciation with the new word prior to the compiling.

14. The language processing module of claim 13 , wherein the identifying comprises receiving the pronunciation from a user.

15. The language processing module of claim 13 , wherein the identifying comprises: generating a plurality of potential pronunciations based on a spelling of the new word; and;
selecting one of the plurality of potential pronunciations that best matches the at least one word as pronounced by a user.

16. The language processing module of claim 12 , further comprising: associating a translation with the new word.

17. The language processing module of claim 16 , wherein the associating comprises: receiving the translation from a user.

18. The language processing module of claim 16 , wherein the associating comprises: retrieving the translation from a dictionary.

19. The language processing module of claim 12 , further comprising: outputting a new subgrammar to a recognition processor.

20. An interactive dialogue system, the system comprising an input device configured to receive a first signal, wherein the first signal comprises speech; and
a language processing module embodied in one or more non-transitory computer readable storage media, the language processing module configured to cause a computing device of the interactive dialogue system to:
generate a recognition result compiling text corresponding to at least a portion of the received first signal; and
with the recognition result:
identify a new word to add to the vocabulary of the interactive computing system, the vocabulary comprising a compiled component and an uninstantiated placeholder class in a search graph;
instantiate the placeholder class with the new word;
compile the instantiated placeholder class independently of the compiled component of the vocabulary;
receive a second signal, the second signal comprising speech; and
compare the instantiated placeholder class to a portion of the second received signal.

Assignees

• Precoda Kristin
• Franco Horacio
• Zheng Jing
• Frandsen Michael
• Abrash Victor
• Akbacak Murat
• Stolcke Andreas
• Stanford Res Inst Int

Inventors

• Precoda Kristin
• Franco Horacio
• Zheng Jing
• Frandsen Michael
• Abrash Victor
• Akbacak Murat
• Stolcke Andreas

Classifications

• G10L15/06 Primary

  Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice (G10L15/14 takes precedence) · CPC title

• G10L15/193

  Formal grammars, e.g. finite state automata, context free grammars or word networks · CPC title

• G10L15/063

  Training · CPC title

• G10L2015/0635

  updating or merging of old and new templates; Mean values; Weighting · CPC title