An efficient, scalable speech recognition architecture combining topic detection and topic-dependent language modeling is proposed for multi-domain spoken language systems. In the proposed approach, the inferred topic is automatically detected from the user's utterance, and speech recognition is then performed by applying an appropriate topic-dependent language model. This approach enables users to freely switch between domains while maintaining high recognition accuracy. As topic detection is performed on a single utterance, detection errors may occur and propagate through the system. To improve robustness, a hierarchical back-off mechanism is introduced where detailed topic models are applied when topic detection is confident and wider models that cover multiple topics are applied in cases of uncertainty. The performance of the proposed architecture is evaluated when combined with two topic detection methods: unigram likelihood and SVMs (Support Vector Machines). On the ATR Basic Travel Expression Corpus, both methods provide a significant reduction in WER (9.7% and 10.3%, respectively) compared to a single language model system. Furthermore, recognition accuracy is comparable to performing decoding with all topic-dependent models in parallel, while the required computational cost is much reduced.

Multi-domain spoken dialogue systems with high degree of intelligence and domain extensibility have long been desired but difficult to achieve. When the user freely surfs among different topics during the dialogue, it will be very difficult for the system to control the switching of the topics and domains while keeping the dialogue consistent, and decide when and how to take the initiative. This paper presents a distributed agent architecture for multi-domain spoken dialogue systems with high domain extensibility and intelligence. Under this architecture, different spoken dialogue agents (SDA's) handling different domains can be developed independently, and then smoothly cooperate with one another to achieve the user's multiple goals, while a user interface agent (UIA) can access the correct spoken dialogue agent through a domain switching protocol, and carry over the dialogue state and history so as to keep the knowledge processed coherently across different domains.