Method of managing a speech cache

• US 6,741,963 B1
• Filed: 06/21/2000
• Issued: 05/25/2004
• Est. Priority Date: 06/21/2000
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Ruback, Harvey M., VanBuskirk, Ronald E., Hanson, Gary R., Woodward, Steven G., Badt, Daniel E., Wang, Huifang, Guasti, Peter J., Nassiff, Amado, Rodriguez, Edwin A., Smith, Carl A.
• Primary Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US09/598,603
• Time in Patent Office: 1,434 Days
• Field of Search: 704/246, 704/251, 704/252, 704/255, 704/275, 704/257, 704/270, 379/88.01
• US Class Current: 704/270
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/285 Memory allocation or algorithm optimisation to reduce hardware requirements

Abstract

A method and a system for improving recall of speech data in a computer speech system can include a plurality of speech cache management steps including providing a speech cache, receiving a speech system input and identifying a speech event in the received speech system input, the speech event comprising speech data. Subsequently, the speech data can be compared to pre-determined speech cache entry criteria; and, if the speech data meets one of the pre-determined entry criteria, at least one entry can be added to the speech cache, the at least one entry corresponding to the speech data. Additionally, the speech data can be compared to pre-determined speech cache exit criteria; and, if the speech data meets one of the pre-determined exit criteria, at least one entry can be purged from the speech cache, the at least one entry corresponding to the speech data. The entry criteria can include frequently used speech data, recently used speech data, and important speech data. Similarly, the exit criteria can include least frequently used speech data associated with each entry in the speech cache, least recently used speech data associated with each entry in the speech cache and least important speed data associated with each entry in the speech cache.

Claims

1. A method of improving recall of speech data in a computer speech recognition system comprising the steps of:

providing a speech cache;

receiving a speech system input;

identifying a speech event in said received speech system input, said speech event comprising speech data;

comparing said speech data to pre-determined speech cache entry criteria;

if said speech data meets one of said pre-determined entry criteria, adding at least one entry to said speech cache, said at least one entry corresponding to said speech data;

comparing said speech data to pre-determined speech cache exit criteria; and

,

if said speech data meets one of said pre-determined exit criteria, purging at least one entry from said speech cache, said at least one entry corresponding to said speech data.

2. The method according to claim 1, wherein said entry criteria comprises frequently used speech data, recently used speech data, and important speech data.

3. The method according to claim 1, wherein said exit criteria comprises least frequently used speech data associated with each entry in said speech cache, least recently used speech data associated with each entry in said speech cache and least important speed data associated with each entry in said speech cache.

4. The method according to claim 1, further comprising the steps of:

comparing entries in said speech cache with filtering criteria; and

,

sorting said entries according to said filtering criteria.

5. The method according to claim 4, wherein said filtering criteria comprises frequency of use of speech data associated with each entry in said speech cache, least recency of use of speech data associated with each entry in said speech cache, and importance of use of speech data associated with each entry in said speech cache.

6. The method according to claim 1, wherein said speech system input is selected from the group consisting of a system event and a speech event.

7. The method according to claim 6, further comprising the steps of:

establishing a table of system events and corresponding speech cache commands;

responsive to receiving a system event, comparing said received system event to said system events in said table; and

,

if said received system event matches a system event in said table, performing said speech cache command corresponding to said matching system event in said table.

8. The method according to claim 7, wherein said corresponding speech cache commands comprise purge commands and add commands.

9. The method according to claim 1, further comprising the steps of:

establishing a frequency counter for said speech data; and

,

responsive to receiving a speech event, incrementing said frequency counter corresponding to said speech data.

10. The method according to claim 1, wherein said comparing step comprises evaluating said speech system input against user-configurable rules for adding and deleting from said speech cache entries corresponding to said speech data, said rules based on frequency of use of said speech data, recency of use of said speech data and importance of use of said speech data.

11. The method according to claim 1, said comparing step comprising the step of evaluating said speech system input against system configured rules for adding and deleting from said speech cache entries corresponding to said speech data, said rules based on a pre-specified list of speech data.

12. The method according to claim 9, wherein said comparing step comprises evaluating said speech system input against user-configurable rules for adding and deleting from said speech cache entries corresponding to said speech data, said rules based on frequency of use of said speech data, said frequency measured by said frequency counter established for said speech data, said adding performed in response to a frequency indicated by said frequency counter exceeding a pre-determined threshold, said deleting performed in response to a frequency indicated by said frequency counter falling below a pre-determined threshold.

13. The method according to claim 1, wherein said comparing steps further comprises the steps of:

comparing said speech system input against system-specified entry and exit criteria and user-specified entry and exit criteria; and

overriding said comparisons performed against said system-specified entry and exit criteria with said comparisons performed against said user-specified entry and exit criteria.

14. The method according to claim 1, further comprising the steps of:

sensing when said speech cache is full; and

,

responsive to sensing a full speech cache, purging entries from said speech cache according to pre-determined purging criteria.

15. The method according to claim 14, wherein said purging criteria comprises least frequently used speech data associated with each entry in said speech cache, least recently used speech data associated with each entry in said speech cache, and least important speech data associated with each entry in said speech cache.

16. The method according to claim 14, wherein said purging step comprises the steps of:

displaying a list of speech cache entries sorted according to said purging criteria;

accepting confirmation from a user before purging entries in said speech cache selected for purging based on said purging criteria; and

,

in response to receiving said confirmation, purging said selected speech cache entries.

17. The method according to claim 1, further comprising the steps of:

associating expiration data with at least one entry in said speech cache; and

,

purging said associated entries in said speech cache according to said expiration data.

18. The method according to claim 17, wherein said associating step comprises the steps of:

accepting user-specified expiration data; and

,

associating said user-specified expiration data with at least one user specified entry in said speech cache.

19. The method according to claim 1, wherein said speech system is an embedded computer speech system.

20. A speech system comprising:

a speech enabled application, said speech enabled application coupled to a speech recognition engine, said speech enabled application and said speech recognition engine configured to process speech data;

a speech cache for storing entries corresponding to said speech data;

predetermined speech cache entry and exit criteria, said entry criteria specifying rules for adding entries corresponding to said speech data to said speech cache, said exit criteria specifying rules for purging entries corresponding to said speech data to said speech cache; and

,

speech cache logic for comparing said speech data to said pre-determined entry and exit criteria, said speech cache logic adding to said speech cache at least one entry corresponding to speech data meeting said pre-determined entry criteria, said speech cache logic purging from said speech cache at least one entry corresponding to speech data meeting said pre-determined exit criteria.

21. The speech system of claim 20, wherein said speech cache is a circular cache.

22. The speech system of claim 20, wherein said entries in said speech cache comprise speech commands, speech objects, pointers to speech commands and pointers to speech objects.

23. The speech system of claim 20, wherein said entries further comprise at least one entry having corresponding expiration data, said speech cache logic purging said at least one entry having corresponding expiration data according to said expiration data.

24. The speech system of claim 20, wherein said speech cache logic is adapted to receive system events in said speech system.

25. The speech system of claim 24, wherein said speech cache logic further comprises a table of system events and corresponding speech cache commands, said speech cache logic adapted to perform a speech cache command in response to receiving a corresponding system event.

26. The speech system of claim 20, wherein said pre-determined entry and exit criteria comprise a speech cache command, frequency of use of said speech data, recency of use of said speech data and importance of use of said speech data.

27. The speech system of claim 26, wherein said speech cache command comprises an add command and a purge command.

28. The speech system of claim 27, further comprising pre-determined purging criteria, said speech cache logic, in response to receiving said purge command, purging entries in said speech cache according to said purging criteria.

29. The speech system of claim 28, wherein said pre-determined purging criteria comprise frequency of use of said speech data corresponding to said entries in said speech cache, recency of use of said speech data corresponding to said entries in said speech cache and importance of use of said speech data corresponding to said entries in said speech cache.

30. The speech system of claim 28, further comprising:

a display for displaying to a user a list of entries in said speech cache selected for purging based on said purging criteria; and

,

means for confirming said purge command before purging said selected entries.

31. The speech system of claim 20, wherein said speech cache logic further comprises filtering logic for sorting said entries in said cache according to pre-determined filtering criteria.

32. The speech system of claim 31, wherein said filtering criteria comprises frequency of use of said speech data corresponding to said entries, recency of use of said speech data corresponding to said entries and importance of use of said speech data corresponding to said entries.

33. The speech system of claim 26, wherein said speech cache logic further comprises at least one incrementable frequency counter corresponding to particular speech data, said frequency counter incremented in response to said speech cache logic receiving an instance of said particular speech data from said speech system, said frequency of use of said particular speech data indicated by said frequency counter.

34. The speech system of claim 20, wherein said speech system is adapted for use in a vehicle.

35. The speech system of claim 20, wherein said speech system is adapted for use in a vehicle navigation system.

36. A machine readable storage, having stored thereon a computer program having a plurality of code sections executable by a machine for causing the machine to perform the steps of:

receiving a speech system input in a speech cache in a computer speech recognition system;

identifying a speech event in said received speech system input, said speech event comprising speech data;

comparing said speech data to pre-determined speech cache entry criteria;

if said speech data meets one of said pre-determined entry criteria, adding at least one entry to said speech cache, said at least one entry corresponding to said speech data;

comparing said speech data to pre-determined speech cache exit criteria; and

,

if said speech data meets one of said pre-determined exit criteria, purging at least one entry from said speech cache, said at least one entry corresponding to said speech data.

37. The machine readable storage according to claim 36, wherein said entry criteria comprises frequently used speech data, recently used speech data, and important speech data.

38. The machine readable storage according to claim 36, wherein said exit criteria comprises least frequently used speech data associated with each entry in said speech cache, least recently used speech data associated with each entry in said speech cache and least important speed data associated with each entry in said speech cache.

39. The machine readable storage according to claim 36, for further causing the machine to perform the steps of:

comparing entries in said speech cache with filtering criteria; and

,

sorting said entries according to said filtering criteria.

40. The machine readable storage according to claim 39, wherein said filtering criteria comprises frequency of use of speech data associated with each entry in said speech cache, least recency of use of speech data associated with each entry in said speech cache, and importance of use of speech data associated with each entry in said speech cache.

41. The machine readable storage according to claim 36, wherein said speech system input is selected from the group consisting of a system event and a speech event.

42. The machine readable storage according to claim 41, for further causing the machine to perform the steps of:

establishing a table of system events and corresponding speech cache commands;

responsive to receiving a system event, comparing said received system event to said system events in said table; and

,

if said received system event matches a system event in said table, performing said speech cache command corresponding to said matching system event in said table.

43. The machine readable storage according to claim 42, wherein said corresponding speech cache commands comprise purge commands and add commands.

44. The machine readable storage according to claim 36, for further causing the machine to perform the steps of:

establishing a frequency counter for said speech data; and

,

responsive to receiving a speech event, incrementing said frequency counter corresponding to said speech data.

45. The machine readable storage according to claim 36, wherein said comparing steps comprises evaluating said speech system input against user-configurable rules for adding and deleting from said speech cache entries corresponding to said speech data, said rules based on frequency of use of said speech data, recency of use of said speech data and importance of use of said speech data.

46. The machine readable storage according to claim 36, said comparing step comprising the step of evaluating said speech system input against system configured rules for adding and deleting from said speech cache entries corresponding to said speech data, said rules based on a pre-specified list of speech data.

47. The machine readable storage according to claim 44, wherein said comparing step comprises evaluating said speech system input against user-configurable rules for adding and deleting from said speech cache entries corresponding to said speech data, said rules based on frequency of use of said speech data, said frequency measured by said frequency counter established for said speech data, said adding performed in response to a frequency indicated by said frequency counter exceeding a pre-determined threshold, said deleting performed in response to a frequency indicated by said frequency counter falling below a pre-determined threshold.

48. The machine readable storage according to claim 36, wherein said comparing step further comprises the steps of:

comparing said speech system input against system-specific entry and exit criteria and user-specified entry and exit criteria; and

overriding said comparisons performed against said system-specified entry and exit criteria with said comparisons performed against said user-specified entry and exit criteria.

49. The machine readable storage according to claim 36, for further causing the machine to perform the steps of:

sensing when said speech cache is full; and

,

responsive to sensing a full speech cache, purging entries from said speech cache according to pre-determined purging criteria.

50. The machine readable storage according to claim 49, wherein said purging criteria comprises least frequently used speech data associated with each entry in said speech cache, least recently used speech data associated with each entry in said speech cache and least important speech data associated with each entry in said speech cache.

51. The machine readable storage according to claim 49, wherein said purging step comprises the steps of:

displaying a list of speech cache entries sorted according to said purging criteria; and

,

accepting confirmation from a user before purging entries in said speech cache selected for purging based on said purging criteria; and

,

in response to receiving said confirmation, purging said selected speech cache entries.

52. The machine readable storage according to claim 36, for further causing the machine to perform the steps of:

associating expiration data with at least one entry in said speech cache; and

,

purging said associated entries in said speech cache according to said expiration data.

53. The machine readable storage according to claim 52, wherein said associating step comprises the steps of:

accepting user-specified expiration data; and

,

associating said user-specified expiration data with at least one user specified entry in said speech cache.

54. The machine readable storage according to claim 36, wherein said speech system is an embedded computer speech system.