Spontaneous Speech Parsing in Travel Information Inquiring and Booking Systems

Grammar-based parsing is a prevalent method for natural languageunderstanding (NLU) and has been introduced into dialogue systems forspoken language processing (SLP). A robust parsing scheme is proposedin this paper to overcome the notorious phenomena, such as garbage,ellipsis, word disordering, fragment, and ill-form, which frequentlyoccur in spoken utterances. Keyword categories are used as terminal symbols, andthe definition of grammar is extended by introducing three new ruletypes, by-passing, up-messing and over-crossing, in additionto the general rules called up-tying in this paper, and the useof semantic items simplifies the semantics extraction. Thecorresponding parser marionette, which is essentially a partialchart parser, is enhanced to parse the semantic grammar. The robustparsing scheme integrating the above methods has been adopted in an airtraveling information service system called EasyFlight, and hasachieved a high performance when used for parsing spontaneous speeches.