The Watson supercomputer that beat past Jeopardy champions was made up of 90 IBM Power 750 Express servers powered by eight-core processors -- four in each machine for a total of 32 processors per machine. The servers were virtualized using a kernel-based virtual machine (KVM) implementation, creating a server cluster with a total processing capacity of 80 teraflops. A teraflop is one trillion operations per second.
The iteration of Watson being used by Cedars-Sinai, which will reside on WellPoint's campus and can be accessed remotely over a WAN, is vastly smaller, according to Gold.
"The Jeopardy! configuration was done with a specific purpose in mind. It was an in-memory application designed to respond to a question in three seconds," Gold said. "It had 2,880 cores and 15 terabytes of memory. Most situations won't dictate that level of response time. For a doctor, if the response is in six seconds or 10 seconds ... obviously the implications for the response are more important than the turnaround time."
IBM's Watson supercomputer as it was used for the game show Jeopardy!. The iteration of Watson being used by Cedars-Sinai will be much smaller, but it will have the same capability to ingest and analyze data from disparate systems, both structured and unstructured
Working with speech and imaging recognition software provider Nuance Communications, IBM said the supercomputer can assist healthcare professionals in culling through gigabytes or terabytes of patient healthcare information to determine how to best treat specific illnesses.
For example, Watson's analytics technology, used with Nuance's voice and clinical language understanding software, could help a physician consider all related texts, reference materials, prior cases, and latest knowledge in journals and medical literature when treating an illness. The analysis could quickly help physicians determine the best options for diagnosis and treatment.
Watson will likely be good at helping physicians prescribe treatments that will have the best outcome, Gold said. For example, between the first and second prescribed treatments of a cancer patient, 50% of the time the prescribed medication changes for the second treatment based on the patient's reaction to the initial treatment, Gold said. Watson may be able to better prescribe initial treatments based on past patient data and information specific to the patient being treated.
"The goal is to assist physicians in evaluating evidence-based treatment options that can be delivered to the physician in a matter of seconds for assessment," he said.
