While IBM’s Watson expert system isn’t ready to take over the world ala Skynet, it’s certainly “smart” enough to beat the world’s best two Jeopardy players. The company isn’t treating this as a trivial exercise; they’re also hard at work turning Watson technology into medical expert systems for cancer research and treatment, as they explained at LinuxCon 2012 in San Diego last month.
Watson, the most famous example of IBM’s DeepQA Project, is made up of ten racks of servers with 15 terabytes of RAM; 2,880 3.55GHz POWER7 processor cores; and a run speed at 80 teraflops. You’re not going to find one of these at your local electronics store.
Watson is far more than an ordinary supercomputer crunching linear equations. No, Watson is meant to solve the far harder problem of “understanding” natural language questions. DeepQA’s eventual goal, according to IBM Fellow David Ferrucci, the principal investigator for Watson technologies, is to create computers that learn through interacting with us. “They will not necessarily require us to sit down and explicitly program them, but through continuous interaction with humans they will start to understand the kind of data and the kind of computation we need,” Ferrucci says.