Dynamic N-best algorithm to reduce recognition errors
View Patent ↗A method for reducing recognition errors. The method includes receiving an N-best list associated with an input of a computer based recognition system. The N-best list includes one or more hypotheses and associated confidence values. The input is classified in response to the N-best list, resulting in a classification. A re-scoring algorithm that is tuned for the classification is selected. The re-scoring algorithm is applied to the N-best list to create a re-scored N-best list. A hypothesis for the value of the input is selected based on the re-scored N-best list.
1. A method for reducing recognition errors, the method comprising:
receiving an N-best list associated with an input of a computer based recognition system, the N-best list including one or more hypotheses and associated confidence values;
classifying the input in response to the N-best list to result in a classification into a bin from a plurality of bins according to characteristics of the N-best list;
selecting a re-scoring algorithm from a plurality of rescoring algorithms that are tuned for the classification, wherein the re-scoring algorithm is selected on a per bin basis;
applying the re-scoring algorithm to the N-best list to create a re-scored N-best list; and
selecting a hypothesis for the value of the input based on the re-scored N-best list.
2. The method of claim 1 wherein:
the computer based recognition system is a speech recognition engine;
the one or more hypotheses and associated confidence values are determined by the speech recognition engine; and
the input is a user utterance.
3. The method of claim 2 wherein the user utterance includes a name of a letter in the alphabet.
4. The method of claim 2 wherein the user utterance includes a name of a number.
5. The method of claim 2 wherein the user utterance includes a word.
6. The method of claim 2 wherein the user utterance includes a phrase.
7. The method of claim 2 wherein the user utterance includes a sentence.
8. The method of claim 1 wherein the computer based recognition system is an optical character reader engine.
9. The method of claim 1 wherein the computer based recognition system is an image recognition engine.
10. The method of claim 1 wherein the re-scoring algorithm was created in response to training data.
11. The method of claim 1 wherein the classifying is based on one or more of the confidence values associated with the one or more hypotheses on the N-best list, an expected frequency of the one or more hypotheses on the N-best list, a conditional probability that the one or more hypotheses are included in the N-best list, confidence value distributions associated with each of the one or more hypotheses, the number of hypotheses on the N-best list, and the order of the hypotheses on the N-best list, where the one or more hypotheses on the N-best list are ordered from highest associated confidence value to lowest associated confidence value.
12. The method of claim 1 wherein the selecting a hypothesis includes selecting the hypothesis with the highest confidence value from the one or more hypotheses on the re-scored N-best list.
13. The method of claim 1 wherein the re-scoring algorithm is based on statistical properties associated with the N-best list.
14. The method of claim 1 wherein the re-scoring algorithm includes re-scoring the N-best list based on the confidence values associated with the one or more hypotheses on the N-best list.
15. The method of claim 1 wherein the re-scoring algorithm includes re-scoring the N-best list based on an expected frequency of the one or more hypotheses on the N-best list.
16. The method of claim 1 wherein the re-scoring algorithm includes re-scoring the N-best list based on a conditional probability that the one or more hypotheses are included in the N-best list.
17. The method of claim 1 wherein the re-scoring algorithm includes re-scoring the N-best list based on confidence value distributions associated with each of the one or more hypotheses.
18. The method of claim 1 wherein the re-scoring algorithm includes re-scoring the N-best list based on the number of hypotheses on the N-best list.
19. The method of claim 1 wherein the re-scoring algorithm includes re-scoring the N-best list based on the order of the hypotheses on the N-best list, where the one or more hypotheses on the N-best list are ordered from highest associated confidence value to lowest associated confidence value.
20. The method of claim 1 wherein the re-scoring algorithm includes one or more of re-scoring the N-best list based on the confidence values associated with the one or more hypotheses on the N-best list, re-scoring the N-best list based on an expected frequency of the one or more hypotheses on the N-best list, re-scoring the N-best list based on a conditional probability that the one or more hypotheses are included in the N-best list, re-scoring the N-best list based on confidence value distributions associated with each of the one or more hypotheses, re-scoring the N-best list based on the number of hypotheses on the N-best list, and re-scoring the N-best list based on the order of the hypotheses on the N-best list, where the one or more hypotheses on the N-best list are ordered from highest associated confidence value to lowest associated confidence value.
21. A computer implemented method for providing a dynamic N-best algorithm to reduce recognition errors, the method comprising:
receiving an N-best list associated with an input of a computer based recognition system, the N-best list including one or more hypotheses and associated confidence values;
classifying the input in response to the N-best list to result in a classification into a bin from a plurality of bins according to characteristics of the N-best list;
selecting a re-scoring algorithm from a plurality of rescoring algorithms that are tuned for the classification, wherein the re-scoring algorithm is selected on a per bin basis;
applying the re-scoring algorithm to the N-best list to create a re-scored N-best list; and
selecting a hypothesis for the value of the input based on the re-scored N-best list.
22. A system for reducing recognition errors, the system comprising a host system in communication with a computer based recognition system, the host system including instructions to implement a method comprising:
receiving an N-best list associated with an input to the computer based recognition system, the N-best list including one or more hypotheses and associated confidence values;
classifying the input in response to the N-best list to result in a classification into a bin from a plurality of bins according to characteristics of the N-best list;
selecting a re-scoring algorithm from a plurality of rescoring algorithms that are tuned for the classification, wherein the re-scoring algorithm is selected on a per bin basis;
applying the re-scoring algorithm to the N-best list to create a re-scored N-best list; and
selecting a hypothesis for the value of the input based on the re-scored N-best list.