Back to results
US 9,262,941 B2Granted
Systems and methods for assessment of non-native speech using vowel space characteristics
AssigneeEducational Testing Services
InventorsLei Chen, Keelan Evanini, Xie Sun
Life of this patent
Estimating the term…
Abstract
Computer-implemented systems and methods are provided for assessing non-native speech proficiency. A non-native speech sample is processed to identify a plurality of vowel sound boundaries in the non-native speech sample. Portions of the non-native speech sample are analyzed within the vowel sound boundaries to extract vowel characteristics. The vowel characteristics are used to identify a plurality of vowel space metrics for the non-native speech sample, and the vowel space metrics are used to determine a non-native speech proficiency score for the non-native speech sample.
Claims44
A computer-implemented method of assessing non-native speech proficiency, comprising:
- processing, using a processing system, a non-native speech sample to identify vowel sounds in the non-native speech sample and vowel sound boundaries corresponding to the vowel sounds in the non-native speech sample;
- analyzing, using the processing system, the vowel sounds within the vowel sound boundaries to extract vowel characteristics;
- processing, using the processing system, the vowel sounds and the vowel characteristics to identify a plurality of vowel space metrics for the non-native speech sample, wherein the vowel space metrics indicate an extent of proficient utilization of AA, IY, and OW vowel sounds in the non-native speech sample;
- and
- generating using the processing system, a non-native speech proficiency score based upon the plurality of vowel space metrics for the non-native speech sample, wherein the non-native speech proficiency score provides a measure of a pronunciation proficiency of the non-native speech sample based upon the plurality of vowel space metrics, wherein the non-native speech proficiency score is generated by processing the plurality of vowel space metrics with a statistical computer model, the statistical computer model comprising multiple weighted features determined by training the statistical computer model relative to a plurality of training items, the training utilizing calculated vowel space metrics of the training items and human-scored non-native speech proficiency scores of the training items to determine values for the weighted features.
2 more elements
A computer-implemented system for assessing non-native speech proficiency, comprising:
- one or more data processors;
- and
- one or more memories, wherein the one or more data processors are configured to execute steps comprising:
- processing a non-native speech sample to identify vowel sounds in the non-native speech sample and vowel sound boundaries corresponding to the vowel sounds in the non-native speech sample;
- analyzing the vowel sounds within the vowel sound boundaries to extract vowel characteristics;
- processing the vowel sounds and the vowel characteristics to identify a plurality of vowel space metrics for the non-native speech sample, wherein the vowel space metrics indicate an extent of proficient utilization of AA, IY, and OW vowel sounds in the non-native speech sample;
- and
- generating a non-native speech proficiency score based upon the plurality of vowel space metrics for the non-native speech sample, wherein the non-native speech proficiency score provides a measure of a pronunciation proficiency of the non-native speech sample based upon the plurality of vowel space metrics, wherein the non-native speech proficiency score is generated by processing the plurality of vowel space metrics with a statistical computer model, the statistical computer model comprising multiple weighted features determined by training the statistical computer model relative to a plurality of training items, the training utilizing calculated vowel space metrics of the training items and human-scored non-native speech proficiency scores of the training items to determine values for the weighted features.
5 more elements
A non-transitory computer-readable storage medium comprising instructions for which when executed cause a processing system to execute steps comprising:
- processing a non-native speech sample to identify vowel sounds in the non-native speech sample and vowel sound boundaries corresponding to the vowel sounds in the non-native speech sample;
- analyzing the vowel sounds within the vowel sound boundaries to extract vowel characteristics;
- processing the vowel sounds and the vowel characteristics to identify a plurality of vowel space metrics for the non-native speech sample, wherein the vowel space metrics indicate an extent of proficient utilization of AA, IY, and OW vowel sounds in the non-native speech sample;
- and
- generating a non-native speech proficiency score based upon the plurality of vowel space metrics for the non-native speech sample, wherein the non-native speech proficiency score provides a measure of a pronunciation proficiency of the non-native speech sample based upon the plurality of vowel space metrics, wherein the non-native speech proficiency score is generated by processing the plurality of vowel space metrics with a statistical computer model, the statistical computer model comprising multiple weighted features determined by training the statistical computer model relative to a plurality of training items, the training utilizing calculated vowel space metrics of the training items and human-scored non-native speech proficiency scores of the training items to determine values for the weighted features.
2 more elements
Dependencies are inferred from the claim text.