Performing actions identified in recognized speech

• US 6,839,669 B1
• Filed: 06/10/1999
• Issued: 01/04/2005
• Est. Priority Date: 11/05/1998
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: ScanSoft, Inc. n/k/a Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Elkins, Michael L., Bayse, Kenneth J., Bamberg, Paul G., Matus, Roger L., Gould, Joel M., Ingold, Charles E., Fieleke, Eric
• Primary Examiner(s): Chawan, Vijay
• Assistant Examiner(s): Opsasnick, Michael N.
• Application Number:
  US09/432,155
• Time in Patent Office: 2,035 Days
• Field of Search: 704/246, 704/249, 704/270, 704/275, 704/276
• US Class Current: 704/246
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A computer is used to perform recorded actions. The computer receives recorded spoken utterances of actions. The computer then performs speech recognition on the recorded spoken utterances to generate texts of the actions. The computer then parses the texts to determine properties of the actions. After parsing the texts, permits the user to indicate that the user has reviewed one or more actions. The computer then automatically carries out the actions indicated as having been reviewed by the user.

Claims

1. A computer-implemented method for performing recorded actions, the method comprising:

at a computer, receiving recorded spoken utterances identifying actions;

performing speech recognition on at least two of the recorded spoken utterances to generate texts identifying the actions;

parsing the texts to determine properties of the actions;

permitting a user to indicate that the user has reviewed properties of one or more actions; and

automatically carrying out the one or more actions indicated as having been reviewed by the user.

2. The method of claim 1, further comprising recording spoken utterances to produce the recorded spoken utterances.

3. The method of claim 2, wherein:

recording the spoken utterances comprises using a portable recorder, and

receiving the recorded spoken utterances comprises transferring the recorded spoken utterances from the portable recorder to the computer.

4. The method of claim 3, wherein the portable recorder comprises a digital recorder and recording a spoken utterance comprises generating a digital data file.

5. The method of claim 3, further comprising using the portable recorder to generate a time stamp indicative of a date and time at which the recorded spoken utterance was produced.

6. The method of claim 5, further comprising transferring time stamps associated with the recorded spoken utterances from the portable recorder to the computer.

7. The method of claim 6, wherein parsing the texts to determine properties of the actions comprises using the associated time stamps.

8. The method of claim 7, wherein, when a text includes a relative date, parsing the text comprises evaluating the relative date relative to the date of the time stamp associated with the action.

9. The method of claim 3, wherein the portable recorder comprises an analog recorder.

10. The method of claim 2, wherein performing speech recognition comprises performing speech recognition on a recorded spoken utterance immediately after the utterance has been spoken.

11. The method of claim 2, wherein recording a spoken utterance comprises using a telephone.

12. The method of claim 1, wherein performing speech recognition comprises performing speech recognition on multiple recorded spoken utterances without user interaction.

13. The method of claim 1, wherein, when the text of an action includes a relative date, parsing the text comprises evaluating the relative date relative to a current date.

14. The method of claim 1, wherein one property of an action comprises an action type.

15. The method of claim 14, further comprising storing an electronic calendar in the computer, wherein carrying out an action for which the action type is an appointment comprises scheduling an appointment in the electronic calendar.

16. The method of claim 15, wherein:

another property of an action is a list of one or more contacts associated with the action, and

scheduling the appointment comprises identifying the list of contacts in the electronic calendar.

17. The method of claim 16, wherein:

another property of an action is that an e-mail confirmation should be sent, and

carrying out an action for which the action type is an appointment comprises sending an e-mail confirmation of the appointment to the associated list of one or more contacts.

18. The method of claim 1, wherein the receiving, performing, and parsing steps are performed for multiple actions in response to a single command from the user and without further user intervention.

19. The method of claim 18, wherein the carrying out step is performed for multiple actions in response to a single command from the user and without further user intervention.

20. The method of claim 1, wherein the carrying out step is performed for multiple actions in response to a single command from the user and without further user intervention.

21. The method of claim 1, further comprising saving unreviewed actions for later review.

22. The method of claim 1, further comprising, after parsing the texts for multiple actions, visually displaying the properties of the actions for review by the user.

23. The method of claim 22, further comprising changing a way in which an action item is displayed when the user indicates that the action item has been reviewed.

24. The method of claim 23, wherein changing the way in which the action item is displayed comprises changing a color in which the action item is displayed.

25. The method of claim 1, wherein a text includes multiple words and parsing the text comprises:

performing a syntactic parsing of the text to assign different words to different properties of the action; and

performing a semantic parsing of the words assigned to each property of the action to generate a value for the associated property.

26. The method of claim 25, wherein performing the syntactic parsing of the text comprises evaluating the text relative to a set of patterns representative of different types of actions.

27. The method of claim 26, further comprising permitting a user to force an action to be of a particular type by stating a designated word at or near the beginning of a spoken utterance for the action.

28. The method of claim 27, wherein parsing the text further comprises:

looking for the designated word at or near the beginning of the text; and

when the designated word is found, performing the syntactic parsing using only a subset of the set of command patterns, the subset including only command patterns associated with the action type corresponding to the designated word.

29. The method of claim 26, wherein a pattern includes multiple elements corresponding to the different properties of an action.

30. The method of claim 25, wherein performing the syntactic parsing of the text comprises using a priority queue.

31. The method of claim 30, wherein performing the syntactic parsing of the text comprises stopping evaluation of a pattern when a best-scoring entry in the priority queue for the pattern does not score as well as a previously-considered pattern.

32. The method of claim 30, wherein performing the syntactic parsing of the text comprises evaluating all of the patterns simultaneously using a single priority queue.

33. The method of claim 30, wherein performing the syntactic parsing of the text comprises evaluating all of the patterns simultaneously using a single priority queue, finding all patterns that share a common prefix, and scoring the prefix portion of only one of the patterns.

34. The method of claim 1, further comprising permitting a user to force an action to be of a particular type by stating a designated word at or near the beginning of a spoken utterance for the action.

35. A computer-implemented method for performing a dictated action, the method comprising:

at a computer, receiving from a communications device a recorded spoken utterance identifying an action;

performing speech recognition on the recorded spoken utterance to generate text identifying the action;

parsing the text to determine properties of the action;

providing an audible indication of the properties of the action using the communications device;

permitting a user to use the communications device to indicate that the user has reviewed properties of the action; and

automatically carrying out the action after the user indicates that the user has reviewed properties of the action.

36. The method of claim 35, further comprising permitting the user to correct properties of the action using the communications device.

37. The method of claim 36, further comprising permitting the user to correct properties of the action using spoken commands.

38. The method of claim 36, wherein the communications device includes keys, the method further comprising permitting the user to correct properties of the action using the keys.

39. The method of claim 35, wherein the communications device comprises a telephone.

40. Computer software, residing on a computer-readable storage medium, comprising instructions for causing a computer to:

receive recorded spoken utterances identifying actions;

perform speech recognition on at least two of the recorded spoken utterances to generate texts identifying the actions;

parse the texts to determine properties of the actions;

permit a user to indicate that the user has reviewed properties of one or more actions; and

automatically carry out the actions indicated as having been reviewed by the user.

41. The computer software of claim 40, further comprising instructions for causing the computer to transfer the recorded spoken utterances from a portable recorder to the computer.

42. The computer software of claim 41, further comprising instructions for causing the computer to transfer date and time stamps associated with the recorded spoken utterances from the portable recorder to the computer.

43. The computer software of claim 42, wherein the instructions for causing the computer to parse the texts to determine properties of the actions use the associated time stamps.

44. The computer software of claim 43, wherein, when the text of an action includes a relative date, the instructions for parsing the text comprise instructions for evaluating the relative date relative to the date of the time stamp associated with the action.

45. A system for performing recorded actions, comprising:

an input device for receiving recorded spoken utterances of multiple actions; and

a processor configured to;

perform speech recognition on at least two of the recorded spoken utterances to generate texts identifying the actions;

parse the texts to determine properties of the actions;

permit a user to indicate that the user has reviewed properties of one or more actions; and

automatically carry out the actions indicated as having been reviewed by the user.