We describe a large-scale investigation of dependency grammar language models. Our work includes several significant departures from earlier studies, notably a larger training corpus, improved model structure, different feature types, new feature selection methods, andmore coherent training and test data. We report word error rate (WER) results of a speech recognition experiment, in which we used these models to rescore the output of the IBM speech recognition system.
Cite as: Berger, A., Printz, H. (1998) Recognition performance of a large-scale dependency grammar language model. Proc. 5th International Conference on Spoken Language Processing (ICSLP 1998), paper 0679, doi: 10.21437/ICSLP.1998-663
@inproceedings{berger98_icslp,
author={Adam Berger and Harry Printz},
title={{Recognition performance of a large-scale dependency grammar language model}},
year=1998,
booktitle={Proc. 5th International Conference on Spoken Language Processing (ICSLP 1998)},
pages={paper 0679},
doi={10.21437/ICSLP.1998-663}
}