Skip to content

Our Dear Watson

February 28, 2011

Photo: Flickr user jntolva

I finally got around to watching the Nova special on Watson, the IBM computer that has become a Jeopardy champion.  I’ve got to admit I’ve become a little bit obsessed with Watson, and artificial intelligence, and this notion that we have started to enter an age where science fiction is no longer fiction, but a reality.

The show was pretty fascinating, and provides a brief look into how they got Watson to be able to answer such difficult questions, full of subtle word play and covering such a wide range of topics. Here’s some information that I pulled from not the Nova special, but a video on the IBM website, on how the DeepQA technology that is behind Watson works:

  • We do not anticipate all questions and build databases.
  • We do not attempt to build a formal model of the world.
  • Focus is on reusable NLP [Natural Language Processing] technology for analyzing vast volumes of as-is text [encylopedias, plays, dictionaries, books, etc.; information as it already exists].  Structured sources (DBs and KBs [databases and knowledge bases]) provide background knowledge for interpreting the text.
  • DeepQA generates and scores many hypotheses using an extensible collection of Natural Language Processing, Machine Learning, and Reasoning Algorithms.  These gather and weigh evidence over both unstructured and structured content to determine the answer with the best confidence.

All the work that went into the creation of this machine completely blows my mind.  I’ve had some people ask me why I’m so fascinated with Watson:  isn’t it just a computer that can play Jeopardy?  But it’s really so much more than that.  The potential applications of this technology are incredible.  Right now, we can question Google, but Google can only point us to sources that may hold potential answers.  We then have to view the material that Google has found and use our minds to determine whether an answer exists within that material.  With technology such as DeepQA, a question would yield an actual answer, not just a set of results.

I found a TED video that discusses the future of Watson and which featured a Columbia University professor who is examining how this technology can be used in the medical field.  The computer would serve as a diagnostic assistant for doctors; it would examine patient history, symptoms, and parse huge databases of medical literature to offer a set of possible diagnoses and probabilities for each.  This is only one of the many potential business applications.

But while I am absolutely amazed by this technology, and am extremely excited to see what developments are made in this field over the next several years, I won’t deny that the potential power of this technology frightens me a bit.  If we are teaching machines to think, does that mean that we will have less reason to use our own minds?  For example, what need is there for a doctor to go to four years of medical school,  if a computer can tell us what a patient’s diagnosis is?  Why spend time reading the latest medical literature if the computer can read it for us?  The Columbia professor on the TED panel said that he could not see having Watson face off against medical residents, as there was no way that the residents could win against Watson.  That is a scary prospect.

Of course, what Watson will lack is the human touch.  Decision makers must be careful to value that which only a flesh and blood doctor can provide.  Abraham Verghese, in a great New York Times op-ed, writes, “An answer that might have been posed on ‘Jeopardy!’ is, ‘An emergency treatment that is administered by ear.’ I wonder if Watson would have known the question (though he will now, cybertroller that he is), which is, ‘What are words of comfort?'”

No comments yet

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: