A fertility channel model for post-correction of continuous speech recognition

Abstract

The authors have implemented a post-processor called SPEECHPP to correct word-level errors committed by an arbitrary speech recognizer. Applying a noisy-channel model, SPEECHPP uses a Viterbi beam-search that employs language and channel models. Previous work demonstrated that a simple word-for-word channel model was sufficient to yield substantial increases in word accuracy. The paper demonstrates that some improvements in word accuracy result from augmenting the channel model with an account of word fertility in the channel. The work further demonstrates that a modem continuous speech recognizer can be used in "black-box" fashion for robustly recognizing speech for which the recognizer was not originally trained. The work also demonstrates that in the case where the recognizer can be tuned to the new task, environment, or speaker, the post-processor can also contribute to performance improvements.