12th Annual Conference of the International Speech Communication Association

Florence, Italy
August 27-31. 2011

Unsupervised Testing Strategies for ASR

Brian Strope, Doug Beeferman, Alexander Gruenstein, Xin Lei

Google Inc., USA

This paper describes unsupervised strategies for estimating relative accuracy differences between acoustic models or language models used for automatic speech recognition. To test acoustic models, the approach extends ideas used for unsupervised discriminative training to include a more explicit validation on held out data. To test language models, we use a dual interpretation of the same process, this time allowing us to measure differences by exploiting expected 'truth gradients' between strong and weak acoustic models. The paper shows correlations between supervised and unsupervised measures across a range of acoustic model and language model variations. We also use unsupervised tests to assess the non-stationary nature of mobile speech input.

Full Paper

Bibliographic reference.  Strope, Brian / Beeferman, Doug / Gruenstein, Alexander / Lei, Xin (2011): "Unsupervised testing strategies for ASR", In INTERSPEECH-2011, 1685-1688.