The Dialogue group works on fundamental research related to the modeling of human-machine communication. Our aim is to improve the naturalness of human-machine interaction by enabling people to converse more easily with machines using speech.
Spoken dialogue systems include a wide range of research fields starting from speech recognition, natural language understanding, dialogue modeling, up to natural language generation and speech synthesis. The Cambridge Research Laboratory has strong experts in many of these fields and we have strong links with the Dialogue Systems Group group at Cambridge University and with other universities.
Our research team is driven to push beyond the boundaries of traditional spoken dialogue systems and explore new methods like neural network based models and end-to-end trainable systems.
We work on all aspects of statistical spoken dialogue system. For single domain dialogues, we are focusing on techniques for fast domain expansion. To support complex queries, we are designing the next generation of statistical spoken dialogue systems using multiple domains as one step toward the open domain statistical dialogue systems. Our patented technologies allow us to add new domains in a fast and low cost manner.