Methods for creating a phrase thesaurus
The invention enables creation of grammar networks that can regulate, control, and define the content and scope of human-machine interaction in natural language voice user interfaces (NLVUI). More specifically, the invention concerns a phrase-based modeling of generic structures of verbal interaction and use of these models for the purpose of automating part of the design of such grammar networks.
1. A phrase analysis system comprising:
a phrase database configured to store phrases; and
a phrase analysis computer coupled with the phrase database and configured to:
obtain a corpus comprising a plurality of phrases;
extract phrases from the corpus;
determine a degree of similarity among the phrases based on phrase contexts associated with the phrases and on overlaps among phrase contexts determined by surrounding phrases in the corpus proximate to the phrases; and
cluster the phrases into classes within the phrase database based on the degree of similarity among the phrase.
2. The system of claim 1 , wherein the corpus comprises a text corpus.
3. The system of claim 1 , wherein the corpus comprises a speech corpus.
4. The system of claim 1 , wherein the corpus comprises a conversational corpus.
5. The system of claim 1 , wherein the corpus comprises a machine-readable corpus.
6. The system of claim 1 , wherein the corpus comprises a specialized corpus that is specific to a domain.
7. The system of claim 6 , wherein the domain comprises an interactive discourse.
8. The system of claim 7 , wherein the interactive discourse comprises at least one of the following: a task oriented interactive discourse, a verbal discourse, a simulated discourse, and a scripted discourse.
9. The system of claim 1 , wherein the classes comprises equivalence classes.
10. The system of claim 1 , wherein the classes comprises tags denoting abstract conceptual representations.
11. The system of claim 10 , wherein the classes comprises hand-checked tags.
12. The system of claim 10 , wherein the tags includes at least one of the following: a request, a confirmation, a rejection, a clarification, an object name, a location, and a proper name.
13. The system of claim 1 , further comprising a dialog manager configured to generate a response to a user based on a user-based natural language input from the user, the response generated from phrases obtained from the phrase database.
14. The system of claim 13 , wherein the user-based natural language input is obtained via a phone.
15. The system of claim 1 , where the phrases comprises normalized phrases.
16. The system of claim 1 , where the corpus reflect interactions in multiple languages.
17. The system of claim 1 , wherein the phrases relate to a genre.
18. The system of claim 1 , wherein the phrases relate to food.
19. The system of claim 1 , wherein the phrases relate to a request for help.