Join the 85,000 open source advocates who receive our giveaway alerts and article roundups.
Could Watson be your doctor's new AI ally? | Opensource.com
Could Watson be your doctor's new AI ally?
Get the newsletter
Watson, IBM’s formidable supercomputer Jeopardy contestant, made a video appearance at the annual meeting of the Healthcare Information and Management Systems Society (HIMSS). Watson appeared in a talk given by Dr. Martin Kohn, chief medical scientist for IBM's Care Delivery Systems.
Kohn explained that Watson incorporates a technology, Deep Question and Answer (DeepQA), that allows it to quickly search through vast amounts of unstructured data (data presented in natural human language), process it, and analyze it in order to respond with an answer the way a human brain can.
Kohn says, “80% of what we live with is unstructured data. It’s how people communicate.” The breakthrough for Watson is in understanding this natural language. It can understand arcane natural language such as puns, anagrams, abbreviations, and rhymes that humans can intuit--but up until now computers could not.
And Watson can process this unstructured natural human language quickly--the equivalent of about a million books or approximately 200 million pages of data in less than three seconds.
So what can Watson’s technology mean to the field of medicine?
“What other area of human endeavor uses intentionally obscure arcane language to inhibit understanding than healthcare?” quipped Dr. Kohn.
The team at IBM believes that this technology can be used to quickly process hundreds (or thousands) of symptoms to aid in medical diagnoses and potentially help treat patients. It was suggested that Watson could help to alert clinicians to adverse drug reactions, form postoperative discharge and follow-up plans, and help manage chronic conditions. Essentially, Watson can fill the “gap between physicians’ knowledge processing capacity and their knowledge processing requirements.” This will be critical, considering that 1.8 zetabytes of data--primarily unstructured--are created annually. At this pace, the amount of medical information doubles every five years.
One of the major challenges in healthcare is incorporating all the data sources Watson could potentially use. This means overcoming the complexities of interacting with existing silos of health data. If Watson can integrate this siloed data, it could present contextually relevant information and evidence-based medicine in a meaningful way to clinicians at the point of care.
And your doctors would have a new AI ally to help improve and speed your medical care.