A Pruned Rnnlm Lattice-Rescoring Algorithm for Automatic Speech Recognition
Hainan Xu, Tongfei Chen, Dongji Gao, Yiming Wang, Ke Li, Nagendra Goel, Yishay Carmiel, Daniel Povey
Abstract
Lattice-rescoring is a common approach to take advantage of recurrent neural language models in ASR, where a word-lattice is generated from 1st-pass decoding and the lattice is then rescored with a neural model, and an <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> -gram approximation method is usually adopted to limit the search space. In this work, we describe a pruned lattice-rescoring algorithm for ASR, improving the n-gram approximation method. The pruned algorithm further limits the search space and uses heuristic search to pick better histories when expanding the lattice. Experiments show that the proposed algorithm achieves better ASR accuracies while running much faster than the standard algorithm. In particular, it brings a 4x speedup for lattice-rescoring with 4-gram approximation while giving better recognition accuracies than the standard algorithm.
BibTeX
@inproceedings{icassp2018_aprunedrnnlmlatt,
title = {A Pruned Rnnlm Lattice-Rescoring Algorithm for Automatic Speech Recognition},
author = {Hainan Xu and Tongfei Chen and Dongji Gao and Yiming Wang and Ke Li and Nagendra Goel and Yishay Carmiel and Daniel Povey and Sanjeev Khudanpur},
booktitle = {ICASSP 2018},
year = {2018}
}