US5712957A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Locating and correcting erroneously recognized portions of utterances by rescoring based on two n-best lists
Assignee: Carnegie Mellon University
•
Filing: 1995-09-08
•
Granted: 1998-01-27
•
Expires: 2015-09-08
Abstract
Abstract A method of repairing machine-recognized speech is comprised of the steps of receiving from a recognition engine a first n-best list of hypotheses and scores for each hypothesis generated in response to a primary utterance to be recognized. An error within the hypothesis having the highest score is located. Control signals are generated from the first n-best list which are input to the recognition engine to constrain the generation of a second n-best list of hypotheses, and scores for each hypothesis, in response to an event independent of the primary utterance. The scores for the hypotheses in the first n-best list are combined with the scores for the hypotheses in the second n-best list. The hypothesis having the highest combined score is selected as the replacement for the located error.
Inventors
- 👤 Alex H. Waibel
- 👤 Arthur E. McNair
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-28 18:44
CPC Classifications
G; G10; G10L; G10L15/00; G10L15/08; G06; G06F; G06F18/00; G06F18/20; G06F18/25; G06F18/254; G06F18/256; G06V; G06V10/00; G06V10/98; G10L15/22; G10L2015/226
Claims Analysis
Total Claims: 17
Claim 1
Independent Claim
156 words
A method of repairing speech recognized by a recognition engine of the type which generates an n-best list of hypotheses and scores for each hypothesis in response to the speech to be recognized, said method comprising the steps of: receiving from a recognition engine a first n-best list of hypotheses, and scores for each hypothesis, generated in response to a primary utterance to be recognized; locating an error within the hypothesis having the highest score: generating control signals from said first n-best list; inputting said control signals to said recognition engine; generating a second n-best list of hypotheses, and scores for each hypothesis, from an event independent of the primary utterance in accordance with said control signals; combining the scores for the hypotheses in the first n-best list with the scores for the hypotheses in the second n-best list; selecting the hypothesis having the highest combined score; and replacing the located error with the selected hypothesis.
Claim 2
Independent Claim
156 words
A method of repairing speech recognized by a recognition engine of the type which generates an n-best list of hypotheses and scores for each hypothesis in response to the speech to be recognized, said method comprising the steps of: receiving from a recognition engine a first n-best list of hypotheses, and scores for each hypothesis, generated in response to a primary utterance to be recognized; locating an error within the hypothesis having the highest score: generating control signals from said first n-best list; inputting said control signals to said recognition engine; generating a second n-best list of hypotheses, and scores for each hypothesis, from an event independent of the primary utterance in accordance with said control signals; combining the scores for the hypotheses in the first n-best list with the scores for the hypotheses in the second n-best list; selecting the hypothesis having the highest combined score; and replacing the located error with the selected hypothesis.
Claim 3
Dependent Claim
36 words
The method of claim 1 additionally comprising the step of displaying the selected hypothesis to a user and, if the selected hypothesis is correct, performing the step of replacing the located error with the selected hypothesis.
Claim 4
Dependent Claim
22 words
The method of claim 2 additionally comprising the step of repeating the method in the event that the selected hypothesis is incorrect.
Claim 5
Dependent Claim
31 words
The method of claim 1 wherein said step of generating a second n-best list includes the step of generating a second n-best list from an event which includes a secondary utterance.
Claim 6
Dependent Claim
21 words
The method of claim 4 wherein said secondary utterance includes a respeaking of at least a portion of the primary utterance.
Claim 7
Dependent Claim
19 words
The method of claim 4 wherein said secondary utterance includes spelling at least a portion of said primary utterance.
Claim 8
Dependent Claim
37 words
The method of claim 1 wherein said step of generating a second n-best list includes the step of generating a second n-best list from an event which includes writing at least a portion of the primary utterance.
Claim 9
Dependent Claim
25 words
The method of claim 1 wherein said step of combining the scores for the hypotheses includes the step of adding the scores for each hypothesis.
Claim 10
Dependent Claim
46 words
The method of claim 1 additionally comprising the step of weighting the hypotheses, and wherein said step of combining the scores for the hypotheses includes the step of adjusting the score for each hypothesis based on the weight assigned to that hypothesis before combining the scores.
Claim 11
Dependent Claim
30 words
The method of claim 1 wherein said step of locating an error within the hypothesis having the highest score includes the step of highlighting the error by using a mouse.
Claim 12
Dependent Claim
30 words
The method of claim 1 wherein said step of locating an error within the hypothesis having the highest score includes the step of highlighting the error by using a keyboard.
Claim 13
Dependent Claim
34 words
The method of claim 1 wherein said step of locating an error within the hypothesis having the highest score includes the step of respeaking the portion of the primary utterance that is in error.
Claim 14
Independent Claim
175 words
A method of repairing speech recognized by a recognition engine of the type which generates an n-best list of hypotheses and scores for each hypothesis in response to the speech to be recognized, said method comprising the steps of: receiving from a recognition engine a first n-best list of hypotheses, and scores for each hypothesis, generated in response to a primary utterance to be recognized; locating an error within the hypothesis having the highest score; parsing each hypothesis in said first n-best list to produced a scored paraphrase for each; receiving from said recognition engine a second n-best list of hypotheses, and scores for each hypothesis, generated in response to a secondary utterance which is a paraphrase of said primary utterance; parsing each hypothesis in said second n-best list to produced a scored paraphrase for each; combining the scores for the paraphrases from the first n-best list with the scores for the paraphrases from the second n-best list; selecting the paraphrase having the highest combined score; and replacing the located error with the selected paraphrase.
Claim 15
Independent Claim
175 words
A method of repairing speech recognized by a recognition engine of the type which generates an n-best list of hypotheses and scores for each hypothesis in response to the speech to be recognized, said method comprising the steps of: receiving from a recognition engine a first n-best list of hypotheses, and scores for each hypothesis, generated in response to a primary utterance to be recognized; locating an error within the hypothesis having the highest score; parsing each hypothesis in said first n-best list to produced a scored paraphrase for each; receiving from said recognition engine a second n-best list of hypotheses, and scores for each hypothesis, generated in response to a secondary utterance which is a paraphrase of said primary utterance; parsing each hypothesis in said second n-best list to produced a scored paraphrase for each; combining the scores for the paraphrases from the first n-best list with the scores for the paraphrases from the second n-best list; selecting the paraphrase having the highest combined score; and replacing the located error with the selected paraphrase.
Claim 16
Dependent Claim
183 words
A speech recognition and repair apparatus for use in a speech recognition system of the type including a recognition engine which generates an n-best list of hypotheses and scores for each hypothesis in response to the speech to be recognized, said apparatus comprising: means for receiving from a recognition engine a first n-best list of hypotheses, and scores for each hypothesis, generated in response to a primary utterance to be recognized; means for locating an error within the hypothesis having the best score; means for generating control signals from said first n-best list, said control signals being input to the recognition engine to constrain the selection of hypotheses; means for receiving from the recognition engine a second n-best list of hypotheses, and scores for each hypothesis, generated in response to an event independent of the primary utterance and in accordance with said control signals; means for combining the scores for the hypotheses in the first n-best list with the scores for the hypotheses in the second n-best list; and means for replacing the located error with the hypothesis having the best combined score.
Claim 17
Dependent Claim
183 words
A speech recognition and repair apparatus for use in a speech recognition system of the type including a recognition engine which generates an n-best list of hypotheses and scores for each hypothesis in response to the speech to be recognized, said apparatus comprising: means for receiving from a recognition engine a first n-best list of hypotheses, and scores for each hypothesis, generated in response to a primary utterance to be recognized; means for locating an error within the hypothesis having the best score; means for generating control signals from said first n-best list, said control signals being input to the recognition engine to constrain the selection of hypotheses; means for receiving from the recognition engine a second n-best list of hypotheses, and scores for each hypothesis, generated in response to an event independent of the primary utterance and in accordance with said control signals; means for combining the scores for the hypotheses in the first n-best list with the scores for the hypotheses in the second n-best list; and means for replacing the located error with the hypothesis having the best combined score.
References Cited (Backward Citations)
US5241619A
Cited
US5579436A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2001-01-26 | 2001-07-28 | 2002-01-27 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2005-01-26 | 2005-07-28 | 2006-01-27 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2009-01-26 | 2009-07-28 | 2010-01-27 | $7700.00 | Window Closed / Lapsed |