Jun 20, 2012
Qualcomm products mentioned within this post are offered by Qualcomm Technologies, Inc. and/or its subsidiaries.
Dissertations on statistical theory may not be at the top of your summer reading list, but one that may actually interest you is the paper that tells you how to live longer.
According to research from the folks at MIT, Columbia, and the University of Washington, our lives might be extended –- or even saved-- thanks to supercomputers capable of running predictive algorithms.
It works like this: when a doctor throws a patient’s medical history into a large data pool of patients, patterns and rules emerge based on the large sample. These rules can be used to make future medical predictions.
PopSci put it best when it likened the principle to the way Netflix or Amazon make viewing suggestions based on your personal interests and others who share similar likes. Similarly, the algorithm predicts future health based on past health and data from a large group.
While the science of prediction is still in the early stages, the concept of crunching numbers to aid in medical decision-making is an idea whose time has come. Until recently, computing power has been a hurdle. But if you ask IBM’s “Watson” supercomputer a question, it can not only sift through 200 million pages of data and provide an answer in under three seconds, but it also processes “natural” language (the way humans speak it).
But what does all this computing mean besides the ability to school humans in Jeopardy? It means that Watson is qualified to work with your future doctor. In a CNNMoney article, Chief Wellpoint Medical Officer, Dr. Sam Nussbaum described Watson saying, “Imagine having the ability to take in all the information about a patient’s medical care…Then imagine using Watson [to analyze]…all of the prior cases…[and] clinical knowledge in the medical literature…to help a physician advance a diagnosis and guide a course of treatment.”
In March of this year, Watson signed up for duty with Memorial Sloan-Kettering Cancer Center (MSKCC) to help oncologists diagnose and treat cancer. If any group could benefit from supercomputers and predictive algorithms, it’s oncologists. As Medical News Today explains, cancer is not one disease, but hundreds of sub-types, each with a different genetic fingerprint, meaning that doctors face a “mind-boggling array of options to review when considering what is best for individual patients.
Watson’s 10 racks of 90 servers with a total of 2,880 processor cores make that job a little easier. And with innovative new algorithms helping identify “rules” for data-based diagnostics, in the not too distant future, predicting the future may not be impossible, after all.