Variational approximation of long-span language models for lvcsr

Long-span language models that capture syntax and semantics are seldom used in the first pass of large vocabulary continuous speech recognition systems due to the prohibitive search-space of sentencehypotheses. Instead, an N-best list of hypotheses is created using tractable n-gram models, and rescored using the long-span models. It is shown in this paper that computationally tractable variational approximations of the long-span models are a better choice than standard n-gram models for first pass decoding. They not only result in a better first pass output, but also produce a lattice with a lower oracle word error rate, and rescoring the N-best list from such lattices with the long-span models requires a smaller N to attain the same accuracy. Empirical results on the WSJ, MIT Lectures, NIST 2007 Meeting Recognition and NIST 2001 Conversational Telephone Recognition data sets are presented to support these claims.

• Type: Journal
• Year: 2011
• Where: ICASSP
• Authors: Anoop Deoras, Tomas Mikolov, Stefan Kombrink, Martin Karafiát, Sanjeev Khudanpur
• Keywords: ICASSP 2011, Long-span Models, N-gram Models, Signal Processing, first Pass