EUROSPEECH 2003 - INTERSPEECH 2003 8th European Conference on Speech Communication and Technology Geneva, Switzerland September 1-4, 2003

Multi-Channel Sentence Classification for Spoken Dialogue Language Modeling

Frédéric Bechet (1), Giuseppe Riccardi (2), Dilek Z. Hakkani-Tur (2)

(1) LIA-CNRS, France
(2) AT&T Labs-Research, USA

In traditional language modeling word prediction is based on the local context (e.g. n-gram). In spoken dialog, language statistics are affected by the multidimensional structure of the human-machine interaction. In this paper we investigate the statistical dependencies of users' responses with respect to the system's and user's channel. The system channel components are the prompts' text, dialogue history, dialogue state. The user channel components are the Automatic Speech Recognition (ASR) transcriptions, the semantic classifier output and the sentence length. We describe an algorithm for language model rescoring using users' response classification. The user's response is first mapped into a multidimensional state and the state specific language model is applied for ASR rescoring. We present perplexity and ASR results on the How May I Help You ?^sm 100K spoken dialogs.

Full Paper

Bibliographic reference.  Bechet, Frédéric / Riccardi, Giuseppe / Hakkani-Tur, Dilek Z. (2003): "Multi-channel sentence classification for spoken dialogue language modeling", In EUROSPEECH-2003, 637-640.