Raymond Lau

Trigger-based language models: a maximum entropy approach

Ongoing efforts at adaptive statistical language modeling are described. To extract information from the document history, trigger pairs are used as the basic information-bearing elements. To combine statistical evidence from multiple triggers, the principle of maximum entropy (ME) is used. To combine the trigger-based model with the static model, the latter is absorbed into the ME formalism. Given consistent statistical evidence, a unique ME solution is guaranteed to exist, and an iterative algorithm exists which is guaranteed to converge to it. Among the advantages of this approach are its simplicity, generality, and incremental nature. Among its disadvantages are its computational requirements. The model described here was trained on five million words of Wall Street Journal text. It used some 40000 unigram constraints, 200000 bigram constraints, 200000 trigram constraints, and 60000 trigger constraints. After 13 iterations, it produced a language model whose perplexity was 12% lower than that of a conventional trigram, as measured on independent data.<<ETX>>

WebGALAXY: beyond point and click—a conversational interface to a browser

Abstract This paper presents WebGALAXY, a flexible multi-modal user interface system that allows wide access to selected information on the World Wide Web (WWW) by integrating spoken and typed natural language queries and hypertext navigation. WebALAXY extends our GALAXY spoken language system, a distributed client-server system for retrieving information from online sources through speech and natural language. WebGALAXY supports a spoken user interface via a standard telephone line as well as a graphical user interface via a standard Web browser using either Java/JavaScript or a cgi-bin/forms front end. Natural language understanding is performed by the system and information servers retrieve the requested information from various online resources including WWW servers, Gopher servers and CompuServe. Currently, queries about three domains are supported: weather, air travel, and points of interest around Boston.

ANGIE: a new framework for speech analysis based on morpho-phonological modelling

This paper describes a new system for speech analysis, ANGIE, which characterizes word substructure in terms of a trainable grammar. ANGIE capture morpho-phonemic and phonological phenomena through a hierarchical framework. The terminal categories can be alternately letters or phone units, yielding a reversible letter-to-sound/sound-to-letter system. In conjunction with a segment network and acoustic phone models, the system can produce phonemic-to-phonetic alignments for speech waveforms. For speech recognition, ANGIE uses a one-pass bottom-up best-first search strategy. Evaluated in the ATIS domain, ANGIE achieved a phone error rate of 36%, as compared with 40% achieved with a baseline phone-bigram based recognizer under similar conditions. ANGIE potentially offers many attractive features, including dynamic vocabulary adaptation, as well as a framework for handling unknown words.