Formerly SpringBlog

Thursday, February 24, 2011

Watson to the Rescue

Language. This is one of the problems associated with information retrieval. Peter Morville, in Ambient Findability, discusses how language hinders our ability to find things in the digital world.

Because words are “imprecise, ambiguous, indeterminate, vague, and opaque” and because spelling errors are commonplace, computers don’t know which meaning a user intends. This ambiguity is one of the reasons we may get lost on the Internet. Even though we know what we want, if a search term has multiple meanings or is spelled incorrectly, the software has no way to determine what we want.

Over the years, Google has tried to improve the precision of search results. For instance, website creators insert tags (metadata) into the HTML. But there is still ambiguity.

Morville states, “computers aren’t even close to extracting or understanding or representing meaning. For as long as humans use language to communicate, information retrieval will remain a messy, imperfect business.”

But wait! This book was published in 2005. This is 2011, and now we have Watson!

If you’re not a Jeopardy fan, here’s a bit of background information. Watson is a computer system that (or should I say “who”?) can understand natural language. According to IBM, “Watson represents a leap forward in data analytics and how this technology will impact business and industry.”

Now, we’re a few years away from having Watsons in our homes. He consists of 90 servers at the moment. But the designers at IBM are hoping that he will transform industries such as healthcare, finance, and telecom. For instance, medical records, journals, and so on are written in natural language. Having a computer that can interpret that language could transform the healthcare industry. Joe Jasinski, the program director for healthcare at IBM, thinks of Watson as “turning data into knowledge.” This brings us back to what Morville considers to be “information” (useful, valid data).

It will be interesting to see what changes take place as Watson evolves. And knowing what we do about Moore’s law, these changes may be here sooner than we expect.


  1. This comment has been removed by the author.

  2. I just watched the Healthcare video at the IBM website(after following the "Watson" link) and I found it very interesting. It would be cool to see if Watson could be implemented the way the medical officials and IBM want it to be implemented. The point that is brought up about possibly being able to consider different variables for different patients instead of treating the average patient (as we all know, not everyone is the same) is a nice goal. This is definitely something I'll keep my eye on.
    (btw, the previous comment was deleted by me, not L Bowman, I had noticed a spelling error.)