Efficient Estimation of Maximum Entropy Language Models with TV-gram features: an SRILM extension

摘要

We present an extension to the SRILM toolkit for training maximum entropy language models with iV-gram features. The extension uses a hierarchical parameter estimation procedure [1] for making the training time and memory consumption feasible for moderately large training data (hundreds of millions of words). Experiments on two speech recognition tasks indicate that the models trained with our implementation perform equally to or better than iV-gram models built with interpolated Kneser-Ney discounting.