En poursuivant votre navigation, vous acceptez l'utilisation de cookies destinés à des fins de mesure d'audience, à améliorer la performance de ce site et à vous proposer des services et contenus personnalisés. En savoir plus

X
Academic affairs

L’X gives the first IBM Watson course in France

École Polytechnique was the first institution in France, along with HEC, to offer courses on Watson supercomputer technologies developed by IBM. A dozen students had the opportunity to attend this initiation in February, 2015.

A computer capable of interacting with human language: that is the Watson system developed by IBM. The Watson supercomputer became famous in 2011 for its performance in the TV game show Jeopardy. Its secret: it understands natural language, ie, the actual language that we use, including its puns, ambiguities and irony.

For the first time in France, courses devoted to Watson technologies were given in February at École Polytechnique and HEC. A dozen students of the Ingénieur Polytechnicien cycle were thus trained during a nine-hour tutorial to discover the supercomputer and its possible applications in everyday life.

A system capable of interacting with the human language

According to Dr. Leo Liberti, a computer science professor at École Polytechnique and Research Director at CNRS, "since the pioneering work of Alan Turing in 1940s, communicating with humans like a human being has been one of the foremost challenges for computers". Alan Turing defined the ability of computers to communicate as a statistical property of a game he called "the Imitation Game". "Turing hoped that technological progress would allow him to see, one day, a computer capable of winning the Imitation Game," says Dr. Liberti, who worked at IBM research for two years. "The Watson system developed by IBM, even if it is not yet capable of winning Turing’s game, represents considerable progress since it is able to interact with human language".

Watson technologies are Natural Language Processing (NLP) techniques. "Natural language, i.e. the language spoken by human beings, is very different from the formal language we program computers with," says Dr. Liberti. Watson is an ecosystem which includes a set of different algorithms. "The system evaluates frequencies and repetitions in sentences, and analyzes the composition of sentences as well as the possible syntax and semantics combinations, ranking them by probability of being correct," says the professor. Watson thus translates natural language into formal language understood by computers. One of the Watson modules can leverage this to give appropriate responses to questions formulated by humans.

Applications in medicine, finance and tourism

But what is the use of Watson? This technology can be applied to many areas of everyday life. For example, Watson can help physicians make better clinical decisions. "Watson has analyzed a large amount of data and can be used to help diagnose diseases by means of symptom descriptions given in natural language, apparently with a good success rate," explains Dr. Liberti. Another area covered by Watson is finance. "Numerous financial reports are produced every day in the world. Watson provides a comprehensive view of this mass of information," says the computer science professor. "Some insurance companies might also choose to assign their telephone information services to Watson, at least for the most common questions". Finally, Watson could be used in the tourism sector to inform travelers about destinations.

Last February, a few École Polytechnique students have attended a series of seminars about this new tool. Oliver Wang, a PhD candidate in the computer science laboratory of École Polytechnique, showed the students concrete examples of services using Watson technologies such as an automatic translation service or a Q&A service. Students also discovered Bluemix, which is the IBM online development platform that provides, among other things, some Watson modules in their Beta versions. "Students now have all the tools to start developing applications that use Watson," says the PhD candidate. Loïc Prieur-Drevon, a student in the Ingénieur Polytechnicien cycle, considers that this course was "an opportunity to demystify what is truly behind the machine that can beat humans at the Jeopardy game!" The student was delighted to learn that IBM was improving open access to rapid prototyping tools that can be of use to entrepreneurship projects.

In the forthcoming years, Dr. Liberti hopes to integrate this training in the curriculum of the Ingénieur Polytechnicien cycle.