Dialogue processing in a conversational speech translation system

Attempts at discourse processing of spontaneously spoken dialogue face several difficulties: multiple hypotheses that result from the parser's attempts to make sense of the output from the speech recognizer, ambiguity that results from segmentation of multi-sentence utterances, and cumulative error-errors in the discourse context which cause further errors when subsequent sentences are processed. In this paper we describe how the JANUS multi-lingual speech-to-speech translation system addresses problems that arise in discourse processing of spontaneous speech. We describe our robust parsers, our procedures for segmenting long utterances, and two approaches to discourse processing that attempt to deal with ambiguity and cumulative error.