Apparatus and method for analysis of language model changes
An apparatus, a method, and a machine-readable medium are provided for characterizing differences between two language models. A group of utterances from each of a group of time domains are examined. One of a significant word change or a significant word class change within the plurality of utterances is determined. A first cluster of utterances including a word or a word class corresponding to the one of the significant word change or the significant word class change is generated from the utterances. A second cluster of utterances not including the word or the word class corresponding to the one of the significant word change or the significant word class change is generated from the utterances.
1. A method comprising:
selecting, via a processor, a plurality of language models; and
for each time period of a plurality of time periods:
identifying a first utterance and a second utterance received during each time period, wherein the first utterance was recognized using a first language model of the plurality of language models and the second utterance was recognized using a second language model of the plurality of language models;
identifying, via the processor, distinctions between the first utterance and the second utterance for each of the plurality of time periods, wherein the distinctions are identified by comparing phonemes between the first utterance and the second utterance;
determining, via the processor, when a significant word usage change has occurred within the first language model and the second language model by comparing the distinctions to previously recorded distinctions;
when the significant word usage change is detected, updating the plurality of language models according to the distinctions; and
outputting a list of utterances on a graphical display with analytical information about the distinctions.
2. The method of claim 1 , wherein determining when the significant word usage change has occurred further comprises comparing:
identifying a first group of utterances from a first time of the plurality of time periods, the first group of utterances having been spoken by a particular group of speakers;
identifying a second group of utterances from a second time of the plurality of time periods, the second group of utterances having been spoken by the particular group of speakers;
when a comparison of the first group of utterances to the second group of utterances indicates the distinctions are above a threshold, identifying that a significant word usage change has occurred.
3. The method of claim 2 , wherein determining when the significant word usage change has occurred further comprises comparing the first group of utterances to a third group of utterances spoken by distinct speakers who are not part of the particular group of speakers.
4. The method of claim 1 , wherein the analytical information further comprises a history of the distinctions.
5. The method of claim 1 , wherein the plurality of time periods comprises two non-overlapping months.
6. The method of claim 1 , further comprising:
iteratively performing, until the language models converge: reassigning each of the first utterance and the second utterance to one of a plurality of subpopulations according to a reassignment criterion.
7. A system comprising:
a processor; and
a computer-readable storage medium having instructions stored which, when executed by the processor, cause the processor to perform operations comprising:
selecting a plurality of language models; and
for each time period of a plurality of time periods:
identifying a first utterance and a second utterance received during each time period, wherein the first utterance was recognized using a first language model of the plurality of language models and the second utterance was recognized using a second language model of the plurality of language models;
identifying distinctions between the first utterance and the second utterance for each of the plurality of time periods, wherein the distinctions are identified by comparing phonemes between the first utterance and the second utterance;
determining when a significant word usage change has occurred within the first language model and the second language model by comparing the distinctions to previously recorded distinctions;
when the significant word usage change is detected, updating the plurality of language models according to the distinctions; and
outputting a list of utterances with analytical information about the distinctions.
8. The system of claim 7 , wherein determining when the significant word usage change has occurred further comprises comparing:
identifying a first group of utterances from a first time of the plurality of time periods, the first group of utterances having been spoken by a particular group of speakers;
identifying a second group of utterances from a second time of the plurality of time periods, the second group of utterances having been spoken by the particular group of speakers; and
when a comparison of the first group of utterances to the second group of utterances indicates the distinctions are above a threshold, identifying that a significant word usage change has occurred.
9. The system of claim 8 , wherein determining when the significant word usage change has occurred further comprises comparing the first group of utterances to a third group of utterances spoken by distinct speakers who are not part of the particular group of speakers.
10. The system of claim 8 , wherein the analytical information further comprises a history of the distinctions.
11. The system of claim 7 , wherein the plurality of time periods comprises two non-overlapping months.
12. The system of claim 7 , the computer-readable storage medium having additional instructions stored which, when executed by the processor, cause the processor to perform operations comprising:
iteratively performing, until the language models converge: reassigning each of the first utterance and the second utterance to one of a plurality of subpopulations according to a reassignment criterion.
13. A computer-readable storage device having instructions stored which, when executed by a computing device, cause the computing device to perform operations comprising:
selecting a plurality of language models; and
for each time period of a plurality of time periods:
identifying a first utterance and a second utterance received during each time period, wherein the first utterance was recognized using a first language model of the plurality of language models and the second utterance was recognized using a second language model of the plurality of language models;
identifying distinctions between the first utterance and the second utterance for each of the plurality of time periods, wherein the distinctions are identified by comparing phonemes between the first utterance and the second utterance;
determining when a significant word usage change has occurred within the first language model and the second language model by comparing the distinctions to previously recorded distinctions;
when the significant word usage change is detected, updating the plurality of language models according to the distinctions; and
outputting a list of utterances with analytical information about the distinctions.
14. The computer-readable storage device of claim 13 , wherein determining when the significant word usage change has occurred further comprises comparing:
identifying a first group of utterances from a first time of the plurality of time periods, the first group of utterances having been spoken by a particular group of speakers;
identifying a second group of utterances from a second time of the plurality of time periods, the second group of utterances having been spoken by the particular group of speakers;
when a comparison of the first group of utterances to the second group of utterances indicates the distinctions are above a threshold, identifying that a significant word usage change has occurred.
15. The computer-readable storage device of claim 14 , wherein determining when the significant word usage change has occurred further comprises comparing the first group of utterances to a third group of utterances spoken by distinct speakers who are not part of the particular group of speakers.
16. The computer-readable storage device of claim 14 , wherein the analytical information further comprises a history of the distinctions.
17. The computer-readable storage device of claim 13 , wherein the plurality of time periods comprises two non-overlapping months.