The DeepQA Research Team - Watson's Jeopardy! Challenge
Leveraging years of experience and research in natural language processing, information retrieval, question-answering, machine learning and AI, David Ferrucci led his team to develop DeepQA as a foundation for building Watson and to take on Jeopardy!
You can read an overview of DeepQA and how it was used to build Watson for the Jeopardy! Challenge in this AI Magazine paper. An entire set of other papers with much greater detail is in the works – you can get a preview of their abstracts here.
Watson was just the beginning. DeepQA is intended to extend well beyond Jeopardy! It is a general architecture that can accommodate many existing techniques in IR, NLP and AI and yet be efficiently extended with new algorithms and new content to address not just one problem but adapt to many domains.
In fact, the DeepQA team at IBM has already begun to demonstrate that the technology can be very quickly adapted to function with the same accuracy, confidence and speed over other domains including Medicine and Technical Support.
We are incredibly excited about what we built. We are looking to extend this in many areas and are on the lookout for the best and brightest researchers to join our team.