Abstract | ||
---|---|---|
In this paper we present a pruning algorithm and experimental results for our recently proposed Sparse Non-negative Matrix (SNM) family of language models (LMs). We show that when trained with only n-gram features SNMLM pruning based on a mutual information criterion yields the best known pruned model on the One Billion Word Language Model Benchmark, reducing perplexity with 18% and 57% over Katz and Kneser-Ney LMs, respectively. We also present a method for converting an SNMLM to ARPA back-off format which can be readily used in a single-pass decoder for Automatic Speech Recognition. |
Year | Venue | Keywords |
---|---|---|
2015 | 16TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2015), VOLS 1-5 | sparse non-negative matrix, language modeling, n-grams, pruning |
Field | DocType | Citations |
Pruning algorithm,Perplexity,Pattern recognition,Computer science,Matrix (mathematics),Speech recognition,Mutual information,Artificial intelligence,n-gram,Language model,Pruning | Conference | 1 |
PageRank | References | Authors |
0.37 | 11 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Joris Pelemans | 1 | 20 | 5.53 |
Noam Shazeer | 2 | 1089 | 43.70 |
Ciprian Chelba | 3 | 1055 | 111.19 |