System and method for combining frame and segment level processing, via temporal pooling, for phonetic classification

摘要

Disclosed herein are systems, methods, and non-transitory computer-readable storage media for combining frame and segment level processing, via temporal pooling, for phonetic classification. A frame processor unit receives an input and extracts the time-dependent features from the input. A plurality of pooling interface units generates a plurality of feature vectors based on pooling the time-dependent features and selecting a plurality of time-dependent features according to a plurality of selection strategies. Next, a plurality of segmental classification units generates scores for the feature vectors. Each segmental classification unit (SCU) can be dedicated to a specific pooling interface unit (PIU) to form a PIU-SCU combination. Multiple PIU-SCU combinations can be further combined to form an ensemble of combinations, and the ensemble can be diversified by varying the pooling operations used by the PIU-SCU combinations. Based on the scores, the plurality of segmental classification units selects a class label and returns a result.

主权项

1. A method comprising:
extracting time-dependent features from an input, to yield extracted time-dependent features; selecting a plurality of time-dependent features from the extracted time-dependent features using a plurality of selection strategies, wherein a plurality of pooling interface units select the plurality of time-dependent features based on a weighted average score and on a rectified average score of the extracted time-dependent features; generating a plurality of feature vectors by pooling the plurality of time-dependent features using a plurality of pooling interface units; generating a plurality of scores associated with the plurality of feature vectors; and returning, in response to the input, a class label selected based on the plurality of scores.