Building the team that built “Watson”

When I was an undergraduate, artificial intelligence seemed to be a discipline long on promise and short on results. (Everthing was about “10 years away”, even 20 years later, it was still “10 years away”.) Therefore, I was pleasantly  impressed when IBM’s Watson machine won an episode of the US quiz show Jeopardy in early 2011 against the all-time champions of the show.

David Ferrucci, leader of the Watson artificial intelligence project , has just written an article for the New York Times in which he describes how he pulled together the multidisciplinary research team behind this great achievement. He had to overcome the researchers’ natural inclinations to stick with their own individual projects and to foster an environment in which they collaborated intensively and shared ideas. He won them over with the vision of how they together could achieve something grand that none of them alone would ever manage. He reinforced the team collaboration with what he terms the “war room”, where the researchers from all the different disciplines worked together in the same big office.

The grand result was due to many insignificant breakthroughs, as if each researcher was delivering one tessera for the overall mosaic.

While winning a quiz show is impressive, it’s just a milestone. IBM is already investigating how Watson’s ability to understand language and crunch data can be put to good use to suggest diagnoses and treatment options to doctors. If you would like to know more about this research, visit IBM’s own site about Watson.
Ferrucci has the last word: “In the end, the hero was the team, not any individual member or algorithm.”

Leave a reply