Grammar learning for spoken language understanding

Ye-Yi Wang1, A. Acero1
1Microsoft Research, Redmond, WA, USA

Tóm tắt

Many state-of-the-art conversational systems use semantic-based robust understanding and manually derived grammars, a very time-consuming and error-prone process. This paper describes a machine-aided grammar authoring system that enables a programmer to develop rapidly a high quality grammar for conversational systems. This is achieved with a combination of domain-specific semantics, a library grammar, syntactic constraints and a small number of example sentences that have been semantically annotated. Our experiments show that the learned semantic grammars consistently outperform manually authored grammars, requiring much less authoring load.

Từ khóa

#Natural languages #Robustness #Authoring systems #Programming profession #Libraries #Information systems #Computer errors #Writing #Law #Legal factors

Tài liệu tham khảo

mielet, 1996, Grammatical Inference: Learning Syntax from Sentences, LNAI 10.1007/BFb0054058 10.3115/980691.980790 stolcke, 0, Inducing Probabilistic Grammars by Bayesian Model Merging, Proceedings of the Second International Colloquium on Grammatical Inference and Applications 1994 wang, 2000, A Plan-Based Dialog System with Probabilistic Inference, Proceedings of ICSLP Beijing China wang, 1999, A Robust Parser For Spoken Language Understanding, 10.21437/Eurospeech.1999-455 10.1109/ICASSP.2001.940754 chelba, 2001, Information Extraction Using the Structured Language Model, EMNLP Pittsburgh 10.1109/2.511967 wang, 2001, Robust Spoken Language Understanding in Milad, Proc Eurospeech Aalborg Denmark gavalda, 2000, Growing Semantics Grammar glass, 2001, SPEECHBUILDER: Facilitating Spoken Dialogue System Development, Proc EUROSPEECH Aalborg Denmark 2001 10.1109/TSMC.1975.5409159 10.1109/TSMC.1975.5409159 zue, 0, JUPITER: A Telephone-Based Conver-sationallnterface for Weather Information, IEEE Trans on Speech and Audio Proc 2000, 10.1109/89.817460 10.1109/ICASSP.1991.150352 10.1007/3-540-58473-0