We already know that IBM's Watson supercomputer is smarter than humans — at Jeopardy at least. But is it smarter than the collective knowledge base that is the world's best oncologists and cancer researchers? A team of doctors at the Cedars-Sinai's Samuel Oschin Comprehensive Cancer Institute in Los Angeles are about to find out.
They're teaming up with IBM and WellPoint, a Blue Cross Blue Shield health plan, to develop a series of applications that will crunch massive amounts of data from thousands of different cancer cases from around the world. If you imagine that a single oncologist would only have access to medical journals, consultations with colleagues, and whatever the Internet's saying these days, it's not hard to realize that a supercomputer capable of storing just about all of the world's knowledge would be nice addition to any doctor's toolkit.
Watson started medical school, so to speak, last fall, when it started tackling different conditions with evidence-based medicine. It's been working towards the cancer challenge for months now. "Where Watson really lends itself to solving problems is information rich opportunities and the information is changing constantly and in various forms, structure and unstructured coming from disparate systems," Steve Gold, director of worldwide marketing for IBM Watson Solutions, told PC World. "Healthcare fits that requirement exceptionally well."
It works not unlike a human brain. By honing its cognitive abilities through data-analysis-techniques like machine-learning and natural-language processing. It can then receive new input, a set of symptoms perhaps, and stack against the information of all the other cases in its database.
Dr. Watson and Jeopardy champ Watson are not exactly the same supercomputer. For the game show, IBM scientists jacked up Watson's response time so it could answer questions in three seconds, which meant that it needed 2,880 cores and 15 terabytes of memory. But Dr. Watson can be a little bit more relaxed about things. Gold explained it could take a few more seconds, especially since it added to quality of the data analysis. Nevertheless, a supercomputer isn't going to cure cancer on its own. It does stand the chance of providing patients with better treatment, and since it will live in the cloud, it'll soon be available to doctors all over the world.
At the very least, it's a cool thing for a computer to do. Better than learning bad words, anyway.
Image via Wikimedia