Paper Info
Title
REGULUS: A Generic Multilingual Open Source Platform for Grammar-Based Speech Applications
Abstract
We present an overview of Regulus, an Open Source platform that supports corpus-based derivation of efficient domain-sp ecific speech recognisers from general linguistically motivated unifica tion grammars. We list available Open Source resources, which include compil- ers, resource grammars for various languages, documentation and a development environment. The greater part of the paper presents a series of experiments carried out using a medium-vocabulary medical speech translation application and a corpus of 801 recorded domain utterances, designed to investigate the impact on speech understanding performance of vocabulary size, grammatical coverage, presence or absence of various linguistic features, degree of genera lity of the grammar and use or otherwise of probabilistic weighting in the CFG language model. In terms of task accuracy, the most significa nt factors were the use of probabilistic weighting, the degr ee of generality of the grammar and the inclusion of features which model sortal restrictions. describe how Regulus allows a domain-specific recogniser to be derived from a general UG, and list Open Source re- sources available under Regulus and related projects which can be used to support this process. The rest of the paper focusses on evaluation, and presents experiments which in- vestigate the impact of various factors on the performance of Regulus-derived recognisers.
Year
Venue
Field
2006
LREC
Rule-based machine translation,Computer science,Compiler,Speech recognition,Grammar,Natural language processing,Artificial intelligence,Speech translation,Vocabulary,Language model,Generality,Sortal
DocType
Citations 
PageRank 
Conference
1
0.48
References 
Authors
6
4
Name
Order
Citations
PageRank
Manny Rayner150889.27
Pierrette Bouillon221441.22
Beth Ann Hockey321236.35
Nikos Chatzichrisafis4356.17