News & Events

02.14.13

Watson Goes to Medical School

IBM’s Watson was invented as an intelligent Question Answering system, made famous in 2011 when it competed with (and trounced) two successful human Jeopardy players.  The Jeopardy victory was a dramatic demonstration of the success of the original vision  for the Watson supercomputer:  to understand complex questions in natural human language.  One of the biggest challenges in making Watson into a conversational and knowledgeable tool is the ambiguous nature of conversational language itself.  Though it may seem simple at first, “natural human language” is no simple matter:  in order to successfully compete in Jeopardy, Watson needed a huge range of knowledge, the ability to analyze subtle meaning, pick up on irony, and generally operate in “human terms,” something that had not been previously accomplished by a machine.

Watson’s success as a bastion of information and knowledge is now being expanded through exciting collaborations with medical institutions such as the Cleveland Clinic and the Memorial Sloan-Kettering Cancer Centerin New York City.  Both collaborations were announced last year, and will require Watson to acquire in-depth access to the ever-growing medical knowledge base.  Researchers and doctors hope that Watson will eventually be able to provide assistance in diagnosis with a probabilistic approach.  The hope is that the future Watson, M.D. will have the ability to quickly find and apply the most pertinent medical knowledge to any medical situation, thus enhancing the capabilities of doctors and medical staff.

In addition, as part of a long-time collaboration with IBM, the Cleveland Clinic plans to allow Watson to interact with medical students, piecing together diagnoses and treatments, much like the students.  Students will interact with Watson, “teaching” it by fine-tuning its conclusions and learning from Watson in turn.

What do you think?

How helpful could a supercomputer be in everyday medicine?

Would a supercomputer help in avoiding bias in decision-making?

 

 


Further Reading:

Watson learns slang: http://tech.fortune.cnn.com/2013/01/07/ibm-watson-slang/

Products

Search