Having conquered Jeopardy!, IBM's Watson is heading to medical school.
Cleveland Clinic is working with Big Blue to further develop the system's "deep question answering" abilities, specifically for medicine. The plan is to have the system regularly work with clinicians, nurses and medical students to help connect the dots on medical facts.
If you're unfamiliar, Watson is a supercomputer that doesn't just memorize, but learns. Inference is key to its potential -- hence its star turn on a television game show.
IBM has been pitching Watson to healthcare institutions as an adviser of sorts, analyzing spoken language input to spit out answers that can help in decision-making. To further develop this talent, Cleveland Clinic and IBM are treating Watson like a student that must work in teams through patient case studies, creating hypotheses based on available evidence (such as electronic medical records) and the field's body of research that ultimately leads to a diagnosis and treatment.
Since IBM is positioning Watson as a unique medical resource, it should be interesting to see how medical students interact with the system, too. Do they rely on it for answers, or merely consult it? How much weight do they give their own clinical experience, versus the machine?
The program is expected to help improve Watson's language and domain analysis capabilities for medicine, specifically. (It learns and gets "smarter" over time.) With luck, it will improve diagnoses, too.