EidoScope

The observable examined

Archive for the ‘Artificial Intelligence’ Category

Are humans unique?

leave a comment »

We have all interacted with animals, seen them at zoos, marveled at them (watching Nat Geo documentaries). If you are like me, you may have wondered whether they are programmed automatons (“fixed action patterns”) or alternately have a “rich” inner life (don’t accuse me of anthropomorphism just yet). I have always held they view that while animals are well adapted to survive their niches, they are able to do more than execute programmed actions.  Also, along the way, I have tired of the notion that “man was made in the image of God”, in the sense that humans are unique and special, given our extraordinary “cognitive” capacities. But then we humans have succumbed to “confirmation bias” by only looking at evidence that confirms this view (as in the tremendous achievements ranging from agriculture to space research). But we also need to review our history of violence and catalog our behaviors collectively as a species (witness terrorism, territoriality, in group/out group, biases, etc.,) and it becomes quickly evident that our “mental” apparatus has a lot in common with other species.

Frans de Waal argues, that this is indeed in the case, in his essay “What I learned from tickling apes” that appears in the NYTimes. Well worth a read. As always, I find it informative to also peruse the reader comments .

Jeopardy meets Watson

leave a comment »

Remember Deep Blue, the chess playing computer from IBM, which beat Gary Kasparov. Now IBM is at it again, this time, Watson is the name of the cyborg (developed at a cost of about $30 million). The goal is to defeat two of the greatest human exponents of Jeopardy, the quiz/knowlegde reality show. The duel will be aired on Feb.14, 15 and 16. 

Watson has been sparring the lesser champions of Jeopardy and, so far, been faring with mixed results. See article for details.  No matter, what is remarkable about this achievement is that it is even possible for a computer to play a tight game against humans, given that the clues offered are a mixture of pure facts  to context sensitive cultural events to downright puns.  Watson has some real heavy lifting to do in parsing the clues and narrowing down the choices even if it has brute force access to an extraordinary database. Hats off IBM team!

I can relate to the excitement at IBM around Watson. I tried to take on  a much less daunting task, to algorithmically code the Mastermind game.  It is easy enough to create this game on the computer and make it give feedback, so that you can deduce the answer. However, the opposite, where you make the computer guess the answer (“artificial intelligence”), gives great insight into human “intelligence”. {Note in the original form of the game, you have to find the secret code in eight or less moves}.

More interestingly, what is the real use of building computers like Watson? For a perspective, see David Gelernter’s article. I agree with Gelernter’s observations. Computers with Watson like capability can be the generalist in the age of specialists. What I mean is, science and medicine are so specialized now that it is really difficult for the practitioners to see the big picture, connect the dots, or reach across disciplines to find valuable connections. A Watson like computer, with access to electronic databases that stretches across discplines, can aid humans researchers connect the dots and make important discoveries that could have far reaching benefits. Go Watson!!

Written by asterix98

February 10, 2011 at 5:46 pm

%d bloggers like this: