Watson Day 2: RPI HEARTS WATSON

by Gabrielle DeMarco on February 15, 2011

Via Twitter Last Night:

@snardiz: Just watched part one of the IBM Jeopardy Challenge at #rpi. Mind. Blown. Panel discussion was really awesome, love this school #Watson

@themole171: Watching #Watson at EMPAC was a TON of fun. Can’t wait for the next 2 nights! #rpi #ibm

Give me a moment to slap on my pocket protector because today we all want to fully embrace the geek within! For those of you who couldn’t make it last night, be sure to make it tonight and tomorrow. Often when there is so much build up around a piece of technology, the reveal can be pretty disappointing. This was not the case with Watson!

As last night’s panelist, Rensselaer Professor Jim Hendler, eloquently put it, watching Watson play literally gave me chills. It was impressive, seamless, and even a little bit frightening just how smart the machine truly is. At first it looked like Watson with its friendly spinning avatar was going to plow through the “wetware” of its human competitors, but the night ended with a cliff hanger – a tie between Brad Rutter and Watson going into the end of the first round of play (cue Watson’s uncertain orange glow).

Tonight, we see if Watson can overtake both Rutter and Jennings. The viewing party at EMPAC tonight begins with another discussion on howWatson works, by IBMer (they like to be called this) and RPI (Rensselaerian?) Chris Welty. This will be followed by another panel led by John Kolb,’79, Rensselaer vice president for information systems and technology and chief information officer. Joining Kolb for the second night is Welty as well as another RPI grad (doesn’t it make you proud?) Adam Lally ’98. Rensselaer computer scientists Barbara Cutler and Christopher Carothers round out the panel.

Adam Lally, ’98, an IBM Senior Software Engineer, worked on Watson’s “Algorithms Team,” which designed and implemented the DeepQA architecture that allows Watson to understand and answer questions at Jeopardy! speed. He will bring some good insight into what went in to Watson that allows it to easily intake and spit out (in that happy synthesized voice) an answer.

Carothers researches methods of optimizing the potential and speed of computing systems. In particular, he has executed simulations using more than 100,000 processors. His current research is investigating advanced massively parallel simulation techniques to enable the co-design of next generation exascale supercomputer systems that will have 10′s of millions of processors – that is a whole lot of Watsons! Watson runs a hundred algorithms simultaneously, and Carothers may speak to how that architecture enables Watson to make the best use of its available computing power. 
 
Cutler is heavily involved in data visualization research. In one example of the immersive, interactive three-dimensional digital display technology that she developed, she and her students were able to literally walk through a huge model of the human brain and even a life-size game of Pong (think Star Trek Holodeck). If her work in visualization research were coupled with a computer of Watson’s power and design, we may be able to find connections and insights that now escape the mind of man.

Kolb, who has been at RPI for over 20 years starting as a student, will make a good host. Kolb is no shrinking violet when it comes to technology. An engineer by training, he works daily to manage the sophisticated technology development on a campus as wired and research-driven as Rensselaer. Kolb was also instrumental in the development and growth of CCNI.

The conversation should be interesting!

Register for the tonight and tomorrow and get all the details at http://watson.rpi.edu/.