Lattice Rescoring Strategies for Long Short Term Memory Language Models in Speech Recognition
Abstract
Recurrent neural network (RNN) language models (LMs) and Long Short Term Memory (LSTM) LMs, a variant of RNN LMs, have been shown to outperform traditional N-gram LMs on speech recognition tasks. However, these models are computationally more expensive than N-gram LMs for decoding, and thus, challenging to integrate into speech recognizers. Recent research has proposed the use of lattice-rescoring algorithms using RNNLMs and LSTMLMs as an efficient strategy to integrate these models into a speech recognition system. In this paper, we evaluate existing lattice rescoring algorithms along with new variants on a YouTube speech recognition task. Lattice rescoring using LSTMLMs reduces the word error rate (WER) for this task by 8\% relative to the WER obtained using an N-gram LM.
Keywords
Cite
@article{arxiv.1711.05448,
title = {Lattice Rescoring Strategies for Long Short Term Memory Language Models in Speech Recognition},
author = {Shankar Kumar and Michael Nirschl and Daniel Holtmann-Rice and Hank Liao and Ananda Theertha Suresh and Felix Yu},
journal= {arXiv preprint arXiv:1711.05448},
year = {2017}
}
Comments
Accepted at ASRU 2017