Discriminative n-gram language modeling Export

@article{citeulike:2359579, abstract = {This paper describes discriminative language modeling for a large vocabulary speech recognition task. We contrast two parameter estimation methods: the perceptron algorithm, and a method based on maximizing the regularized conditional log-likelihood. The models are encoded as deterministic weighted finite state automata, and are applied by intersecting the automata with word-lattices that are the output from a baseline recognizer. The perceptron algorithm has the benefit of automatically selecting a relatively small feature set in just a couple of passes over the training data. We describe a method based on regularized likelihood that makes use of the feature set given by the perceptron algorithm, and initialization with the perceptron's weights; this method gives an additional 0.5\% reduction in word error rate (WER) over training with the perceptron alone. The final system achieves a 1.8\% absolute reduction in WER for a baseline first-pass recognition system (from 39.2\% to 37.4\%), and a 0.9\% absolute reduction in WER for a multi-pass recognition system (from 28.9\% to 28.0\%).}, author = {Roark, B. and Saraclar, M. and Collins, M.}, citeulike-article-id = {2359579}, citeulike-linkout-0 = {http://dx.doi.org/10.1016/j.csl.2006.06.006}, citeulike-linkout-1 = {http://linkinghub.elsevier.com/retrieve/pii/S0885230806000271}, citeulike-linkout-2 = {http://www.sciencedirect.com/science/article/B6WCW-4KJV20T-1/2/a68af59be3ad8ab29c2a079c512fcbfc}, doi = {10.1016/j.csl.2006.06.006}, issn = {08852308}, journal = {Computer Speech \& Language}, keywords = {discriminative, lm, ngram}, month = {April}, number = {2}, pages = {373--392}, posted-at = {2009-10-08 01:26:08}, priority = {2}, title = {Discriminative n-gram language modeling}, url = {http://dx.doi.org/10.1016/j.csl.2006.06.006}, volume = {21}, year = {2007} }

TY - JOUR ID - citeulike:2359579 L3 - citeulike-article-id:2359579 N2 - This paper describes discriminative language modeling for a large vocabulary speech recognition task. We contrast two parameter estimation methods: the perceptron algorithm, and a method based on maximizing the regularized conditional log-likelihood. The models are encoded as deterministic weighted finite state automata, and are applied by intersecting the automata with word-lattices that are the output from a baseline recognizer. The perceptron algorithm has the benefit of automatically selecting a relatively small feature set in just a couple of passes over the training data. We describe a method based on regularized likelihood that makes use of the feature set given by the perceptron algorithm, and initialization with the perceptron’s weights; this method gives an additional 0.5% reduction in word error rate (WER) over training with the perceptron alone. The final system achieves a 1.8% absolute reduction in WER for a baseline first-pass recognition system (from 39.2% to 37.4%), and a 0.9% absolute reduction in WER for a multi-pass recognition system (from 28.9% to 28.0%). IS - 2 JF - Computer Speech & Language SN - 08852308 EP - 392 TI - Discriminative n-gram language modeling VL - 21 SP - 373 KW - discriminative KW - lm KW - ngram AU - Roark, B AU - Saraclar, M AU - Collins, M PY - 2007/04// UR - http://dx.doi.org/10.1016/j.csl.2006.06.006 ER -