The paper presents an in-depth analysis of a less known interaction between Kneser-Ney smoothing and entropy pruning that leads to severe degradation in language model performance under aggressive pruning regimes. Experiments in a data-rich setup such as verb+google.com+ voice search show a significant impact in WER as well: pruning Kneser-Ney and Katz models to 0.1% of their original impacts speech recognition accuracy significantly, approx. 10% relative.
Cite as: Chelba, C., Brants, T., Neveitt, W., Xu, P. (2010) Study on interaction between entropy pruning and kneser-ney smoothing. Proc. Interspeech 2010, 2422-2425, doi: 10.21437/Interspeech.2010-525
@inproceedings{chelba10_interspeech, author={Ciprian Chelba and Thorsten Brants and Will Neveitt and Peng Xu}, title={{Study on interaction between entropy pruning and kneser-ney smoothing}}, year=2010, booktitle={Proc. Interspeech 2010}, pages={2422--2425}, doi={10.21437/Interspeech.2010-525} }