In fact, Watson has been downtown learning about law and legal research at the University of Toronto since sometime early this fall. IBM approached 10 schools and challenged them to “put an entrepreneurial spin on Watson’s artificial intelligence.” U of T was the only Canadian university invited to participate in this IBM Watson Cognitive Computing Competition.
IBM is hoping to kick start innovation in cognitive computing by “incorporating the technology into an undergraduate curriculum that combines computing skills with entrepreneurship, and has invited U of T’s computer science department to participate in the program.” This participation took the form of a one-time only course offered this fall called the “Cognitive Computing Option.”
The U of T students decided to focus their efforts on using Watson in the legal domain. They will “… develop various applications that can access vast amounts of legal information. Students are developing the skills to upload relevant legal information into Watson’s body of knowledge and train it to collate this unstructured data to produce relevant responses.”
Yana Davis is a member of one of the U of T teams working on this project and she and her instructor Mario Grech spoke about their experiences on CBC radio a couple of weeks ago. Grech says they’ve been loading in domain specific information and essentially “training Watson like a child that knows nothing and raising the intellectual capabilities of the application.” Davis noted that although search engines retrieve a collection of documents based on keyword input Watson is “taught how to answer questions” and analyse the nuances of natural language.
As a computer science student Davis admits that she and her colleagues don’t “know how a lawyer thinks and works.” So their team has been consulting with legal advisors to help them understand the legal research environment. They agree that Watson is not a “super lawyer” and is not expected to replace lawyers … “at least right now.”
Davis feels that what their application might be able to do is help “predict the outcome of a certain case, based on historical data” using “resources that lawyers [would] commonly use to arrive at the same conclusions.” She imagines it as similar to a “law librarian that knows every single book in the library, every single case that ever existed.”
There were four other teams working with Watson at U of T. Two were working on similar legal research tools, one on an application to help people manage their divorces and the last was working on an application that provides advice on immigration law. Each team is competing to be selected to participate in the larger IBM Watson Challenge competition which takes place in New York next month. The team that provides the “most insightful and articulate business proposal for the IBM Watson platform” will receive a $100,000 prize to seed the further development of their product.
Davis’s team was not selected as the winning U of T team. Jeff Gray confirmed in the Globe and Mail last Friday that the team working on an application called Ross , “the best legal researcher available,” was the team U of T will be sending to compete next month against the nine other universities involved.
Michael Rhodin, senior vice-president of IBM Watson Group says, “By putting Watson in the hands of tomorrow’s innovators, we are unleashing the creativity of the academic community into a fast-growing ecosystem of partners who are building transformative cognitive computing applications.” And it’s also interesting to note that IBM has also released Watson’s analytic platform as a new public beta. As Chris Mills reports on Gizmodo, everyone now has the opportunity to “chuck in a dataset” and let Watson “pull out the interesting correlations, predictive analyses and the like, and present it all in a series of infographics and graphs.”
Looking forward to finding out who will be taking their Watson application and entrepreneurial vision forward in the New Year.