The story of Watson and the Grand Challenge is pure fascination for tech geeks like me. For those of you who don’t know or haven’t heard yet, Watson is the latest supercomputer from IBM that will use Natural Language Processing (NLP) and face off against two former Jeopardy champions on national television in mid-February 2011.
What interests me so much about this concept isn’t the fact that human beings will be pitted against a supercomputer to determine the winner; we’ve seen similar exchanges such as Garry Kasparov versus Deep Blue in the legendary chess matches of the 90s and pretty much every time I start my favorite PC game in computer-player mode. In the case of Watson, I don’t think it really matters isn’t who wins or who loses, but what the event itself foreshadows.
We are witnessing a pivotal moment in history where we’ve gone beyond just the static strategy of technology and are now waist-deep in a realm where machines have to the ability to combine human linguistics against probabilistic matching mixed with self-corrective learning. Think of a search engine with the ability to distinguish human phrasings in unique ways.
In short, we are seeing a form of Artificial Intelligence which is as close to human thought as we have yet been able to achieve.
Watson represents a roadmap for the near future, especially in global fields such as health care, legal advice, and technical support. The concept of going to a physician or lawyer for initial consultation may become quickly outdated as would populating call centers across the world with Tier 1 support staff. We are already headed in this direction with technologies like the cloud and Voice Recognition (VR) and a demonstration of Watson only adds more fuel towards the technological fire.
Before I claim that we’re headed into the realm previously reserved for science fiction fans, let me be perfectly clear: Watson is not the forerunner of a human/machine war and is not intended to replace human interaction at the more complex levels. That is, Watson cannot yet replace a true medical specialist or the intricate levels of jurisprudence designed for lawyers with even a few years of experience.
In its current state, Watson is rather limited. The system cannot understand audio or video stimuli and, as I understand it, receives its input via fast text messages streaming across a network TCP/IP port. In addition, Watson lacks the human quality of anticipation which leads to a critical lag time in response: rather than hearing a question and then responding independently, Watson must be “triggered” with a signal such as an “answer now” command in sync with the final word of the host’s question.
These limitations, as well as others, put Watson somewhat at a disadvantage over its human counterparts. I wouldn’t be greatly surprised if it did not fare as well as expected, but am still looking forward to a great display nonetheless.
The part I most anticipate is the moment when the simulated voice of Watson answers that first question. That’s the moment when we get a glimpse of our future and can feel the rush of excitement that occurs the moment you realize that you are seeing something that can change the world.