An information gain and grammar complexity based approach to attribute selection in speech enabled information retrieval dialogs

Effective dialog driven method is required for today's speech enabled information retrieval systems, such as name dialer. Similar to dynamic sales dialog for electronic commerce scenarios, the information gain measure based approaches are widely used for attribute selection and dialog length reduction. However for speech enabled information retrieval systems, another important factor influencing attribute selection is speech recognition accuracy. Too low accuracy will result in a failed dialog. Recognition accuracy varies with many issues including acoustic model performance, grammar's complexity. Acoustic model is fixed for a whole dialog, while grammar is different for each interaction round, thereby grammar complexity will influence the attribute selected for next question. In mis paper, an approach combining both information gain measurement and grammar complexity is present for dynamic dialog driven. Off-line evaluations show that this approach can give a trade-off of faster discriminating the candidates for retrieval target and higher recognition accuracy. © 2004 IEEE.