For the past three years, IBM scientists have been working on an advanced computing system and have made significant progress toward building a system capable of competing with human contestants at the game of Jeopardy!. The computer -- codenamed "Watson" -- is being designed to quickly answer complex questions with pinpoint precision.
The computing engine is much more than a simple Web search engine. Watson is a sophisticated system that does not just search on keywords, but actually understands idiomatic expressions, puns, linguistics and can parse clues and come up with answers.
Watson is based on the scientific discipline called “The Open Advancement of Question Answering” (OAQA) – an initiative spear-headed by IBM and Carnegie Mellon University to provide a foundation for effective collaboration among researchers to accelerate the science of automatic question answering. The initiative focuses on developing common metrics, architectures, experimental methodologies and tools to help facilitate the collaborative advancement of the state-of-the-art in QA.
Beyond Jeopardy!, the challenge expands to demonstrating how QA technologies can be quickly and effectively adapted to different business applications. These applications demand deep understanding of a user's questions – which may be stated in phrases that have subtle meanings - and the processing of huge volumes of information to rapidly deliver and justify precise and succinct answers or to support intelligent and consultative dialogs.
Building a system that can compete at Jeopardy! is a good surrogate for business and societal applications and will have significant impact in areas such as healthcare, financial services, customer relationship management and environmental issues.
Following are links to the New York Times article from June 20, 2010 and a four minute YouTube on Watson and some of the IBM team that designed it.