Automatic Feature Selection for Predicting Content of User Utterances in Dialogs

In task-oriented spoken dialog systems (SDS), the system often requests explicit confirmation of userprovided task-relevant concepts. The user utterance following a confirmation question (the postconfirmation utterance) is important to successful dialog outcomes. It may be a simple confirmation or rejection (e.g. yes, no, right, correct), or a correction or a topic change containing new concepts. For example, in a set of one month of calls to the deployed Let’s Go! bus route information SDS (Raux et al., 2005), 18% of post-confirmation utterances contain a concept (time, place, or bus) and 20% of these contain a concept type different from that in the system’s confirmation prompt. The speech recognition word error rate (WER) on all post-confirmation utterances in this set is 38%, while on post-confirmation utterances with a concept it is 49%. Correct identification of concept types in post-confirmation utterances could lead to improved speech recognition and dialog outcomes. In this paper, we propose a concept-specific language model adaptation strategy and evaluate it on post-confirmation utterances. We adopt a two-pass recognition approach (Young, 1994). In first pass recognition, the input utterance is processed using a generic language model trained on post-confirmation utterances. Recognition with a generic model frequently fails on concept words such as Oakland or 61C. We then use acoustic, lexical and dialog history features to determine the task-related concept type(s) likely to be present in the utterance. Finally, any utterance that is determined to contain a concept type is re-processed using a concept-specific language model. We show that: (1) it is possible to achieve high accuracy in determining presence or absence of particular concept types in a post-confirmation utterance; and (2) concept-specific language model adaptation can lead to improved speech recognition performance for post-confirmation utterances. In this paper we focus on alternative methods for selecting lexical features for concept type classification in the presence of first-pass recognition errors. 2. Classification Experiment