METHOD AND SYSTEM FOR AUTOMATICALLY DETECTING MORPHEMES IN A TASK CLASSIFICATION SYSTEM USING LATTICES

• US 20080215328A1
• Filed: 09/13/2007
• Published: 09/04/2008
• Est. Priority Date: 11/05/1999
• Status: Abandoned Application

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: AT&T Corporation (AT&T, Inc.)
• Inventors: Riccardi, Giuseppe, Wright, Jeremy Huntley, Gorin, Allen Louis, Petrovska-Delacretaz, Dijana
• Application Number:
  US11/854,706
• Publication Number:
  US 20080215328A1
• Time in Patent Office: Days
• US Class Current: 704/254
• CPC Class Codes: G10L 15/08 Speech classification or search

Abstract

The invention concerns a method and system for detecting morphemes in a user'"'"'s communication. The method may include recognizing a lattice of phone strings from the user'"'"'s input communication, the lattice representing a distribution over the phone strings, and detecting morphemes in the user'"'"'s input communication using the lattice. The morphemes may be acoustic and/or non-acoustic. The morphemes may represent any unit or sub-unit of communication including phones, diphones, phone-phrases, syllables, grammars, words, gestures, tablet strokes, body movements, mouse clicks, etc. The training speech may be verbal, non-verbal, a combination of verbal and non-verbal, or multimodal.

Claims

1. A method for detecting morphemes in a user'"'"'s input communication, comprising:

forming a lattice representing a distribution of acoustic and non-acoustic phone strings associated with morphemes in the user'"'"'s acoustic and non-acoustic input; and

detecting acoustic and non-acoustic morphemes in the formed lattice.

2. The method of claim 1, wherein detecting the acoustic and non-acoustic morphemes further comprises extracting the N-best phone strings from the formal lattice.

3. The method of claim 1, wherein detecting the acoustic and non-acoustic morphemes further comprises extracting the N-best phone strings and their confidence scores from the formal lattice.

4. The method of claim 1, wherein the non-acoustic communication includes use of at least one of gestures, body movements, head movements, non-responses, text, keyboard entries, keypad entries, mouse clicks, DTMF codes, pointers, stylus, cable set-top box entries, graphical user interface entries and touchscreen entries.

5. The method of claim 1, wherein the user'"'"'s input communication is in multimodal form.

6. The method of claim 1, wherein the user'"'"'s acoustic input is verbal.

7. The method of claim 1, wherein detecting acoustic and non-acoustic morphemes is performed using morphemes from a morpheme generation subsystem.

8. The method of claim 1, further comprising:

providing the detected acoustic and non-acoustic morphemes to a task classification processor.

9. The method of claim 8, further comprising:

based on a probabilistic relationship, implementing a task objective between the detected morphemes and selected task objectives.

10. A system for detecting morphemes in a user'"'"'s input communication, the system comprising:

a module configured to form a lattice representing a distribution of acoustic and non-acoustic phone strings associated with morphemes in the user'"'"'s acoustic and non-acoustic input; and

a module configured to detect acoustic and non-acoustic morphemes in the formed lattice.

11. The system of claim 10, wherein detecting the acoustic and non-acoustic morphemes further comprises a module configured to extract the N-best phone strings from the formal lattice.

12. The system of claim 10, wherein detecting the acoustic and non-acoustic morphemes further comprises a module configured to extract the N-best phone strings and their confidence scores from the formal lattice.

13. The system of claim 10, wherein the non-acoustic communication includes use of at least one of gestures, body movements, head movements, non-responses, text, keyboard entries, keypad entries, mouse clicks, DTMF codes, pointers, stylus, cable set-top box entries, graphical user interface entries and touchscreen entries.

14. The system of claim 10, wherein the user'"'"'s input communication is in multimodal form.

15. The system of claim 10, wherein the user'"'"'s acoustic input is verbal.

16. The system of claim 10, wherein detecting acoustic and non-acoustic morphemes is performed using morphemes from a morpheme generation subsystem.

17. The system of claim 10, further comprising:

a module configured to provide the detected acoustic and non-acoustic morphemes to a task classification processor.

18. The system of claim 17, further comprising:

a module configured to implement a task objective between the detected morphemes and selected task objectives and based on a probabilistic relationship.

19. A computer readable medium storing instructions for controlling a computing device to detect morphemes in a user'"'"'s input communication, the instructions comprising:

forming a lattice representing a distribution of acoustic and non-acoustic phone strings associated with morphemes in the user'"'"'s acoustic and non-acoustic input; and

detecting acoustic and non-acoustic morphemes in the formed lattice.

20. The computer readable medium of claim 19, wherein detecting the acoustic and non-acoustic morphemes further comprises extracting the N-best phone strings from the formal lattice.

21. The computer readable medium of claim 19, wherein detecting the acoustic and non-acoustic morphemes further comprises extracting the N-best phone strings and their confidence scores from the formal lattice.

22. The computer readable medium of claim 19, wherein the non-acoustic communication includes use of at least one of gestures, body movements, head movements, non-responses, text, keyboard entries, keypad entries, mouse clicks, DTMF codes, pointers, stylus, cable set-top box entries, graphical user interface entries and touchscreen entries.

23. The computer readable medium of claim 19, wherein the user'"'"'s input communication is in multimodal form.

24. The computer readable medium of claim 19, wherein the user'"'"'s acoustic input is verbal.

25. The computer readable medium of claim 19, wherein detecting acoustic and non-acoustic morphemes is performed using morphemes from a morpheme generation subsystem.