Practical Issues in Compiling Typed Unification Grammars for Speech Recognition

Practical Issues in Compiling Typed Unification Grammars for Speech Recognition John Dowding Beth Ann Hockey RIACS RIALIST Group NASA Ames Research Center Moffett Field, CA 94035 jdowding@riacs.edu bahockey@riacs.edu Jean Mark Gawron Dept. of Linguistics San Diego State University San Diego, CA gawron@mail.sdsu.edu Christopher Culy SRI International 333 Ravenswood Avenue Menlo Park, CA 94025 culy@ai.sri.com Abstract Current alternatives for language mod- eling are statistical techniques based on large amounts of training data, and hand-crafted context-free or finite-state grammars that are difficult to build and maintain. One way to address the problems of the grammar-based ap- proach is to compile recognition gram- mars from grammars written in a more expressive formalism. While theoreti- cally straight-forward, the compilation process can exceed memory and time bounds, and might not always result in accurate and efficient speech recogni- tion. We will describe and evaluate two approaches to this compilation prob- lem. We will also describe and evalu- ate additional techniques to reduce the structural ambiguity of the language model. 1 Introduction Language models to constrain speech recogni- tion are a crucial component of interactive spo- ken language systems. The more varied the lan- guage that must be recognized, the more critical good language modeling becomes. Research in language modeling has heavily favored statisti- cal approaches (Cohen 1995, Ward 1995, Hu et al. 1996, Iyer and Ostendorf 1997, Bellegarda 1999, Stolcke and Shriberg 1996) while hand- coded finite-state or context-free language models dominate the commercial sector (Nuance 2001, SpeechWorks 2001, TellMe 2001, BeVocal 2001, HeyAnita 2001, W3C 2001). The difference re- volves around the availability of data. Research systems can achieve impressive performance us- ing statistical language models trained on large amounts of domain-targeted data, but for many domains sufficient data is not available. Data may be unavailable because the domain has not been explored before, the relevant data may be con- fidential, or the system may be designed to do new functions for which there is no human-human analog interaction. The statistical approach is un- workable in such cases for both the commercial developers and for some research systems (Moore et al. 1997, Rayner et al. 2000, Lemon et al. 2001, Gauthron and Colineau 1999). Even in cases for which there is no impediment to col- lecting data, the expense and time required to col- lect a corpus can be prohibitive. The existence of the ATIS database (Dahl et al. 1994) is no doubt a factor in the popularity of the travel do- main among the research community for exactly this reason. A major problem with grammar-based finite- state or context-free language models is that they can be tedious to build and difficult to maintain, as they can become quite large very quickly as the scope of the grammar increases. One way to address this problem is to write the gram- mar in a more expressive formalism and gener- ate an approximation of this grammar in the for- mat needed by the recognizer. This approach has been used in several systems, CommandTalk (Moore et al. …

Content truncated to 500 words