I recently ran across this interesting video during one of my many sessions with StumbleUpon. The video via GeeksAreSexy (shown after the “more”) demonstrates a guy in a robotic exoskeleton performing a variety of tasks to demonstrate the strength of the device. This was a weird coincidence because I was in the process of reading Starship Troopers by Robert A. Heinlein. His Mobile Infantry armor was much more elaborate, but I’m always interested when things from science fiction start to become science fact.
I was a bit skeptical of the cable hooked up to the back, because that didn’t seem very “portable,” but the video goes on to explain that the final version will be 100% portable with varying levels of armor protection. What is most interesting is that they talk about the goal of making the exoskeleton operate as an autonomous robot when the person steps out of the suit. I’m sure lots of research will have to be completed before the programming of such a robot would be possible.
A few months ago a co-worker of mine suggested that I read The Age of Spiritual Machines: When Computers Exceed Human Intelligence by Ray Kurzweil. We had been talking about AI and he mentioned that this was an interesting read. I wasn’t sure exactly what to expect, as I had never heard of Kurzweil before. After a few Google searches I got the impression that he was a quirky futurist.
This book was published back in 1999 and by the end looks very much like science fiction. Many of his predictions are founded on some sort of research. I can see how they would have seemed a bit “out there” almost 10 years ago. I wondered how many of his predictions would hold up. Central to his philosophy is the Law of Accelerating Returns. In short, technology will continue to progress ever faster as time goes on. He displays an interesting graph of the exponential increase of computational power in various models of computers over time. The line is slightly curved upward, which represents an increase in exponential growth over time. So, according to Kurzweil computation is progressing exponentially exponentially faster. (more…)
For all of you geeks that have had your head in the sand, artificial intelligence has hit a milestone. Yesterday, IBM’s Watson trounced these bags of meat known as Ken Jennings and Brad Rutter in Jeopardy. If I ruined the result of the match, I apologize, but I figure posting a day after is enough notice for anyone that was following this from the start. For some more background on Watson and behind the scenes info, check out my original post on this. I have to admit that I was rooting for Watson from the start. I was a bit worried when after the first day in the tournament Watson was only tied for first place. I’m not sure what happened between the first and second matches, but Watson rocketed ahead in the second day and never looked back.
I think IBM might have been in a rush to show off their new creation. It was interesting to see the answer confidence levels during the rounds, often revealing some really wacky possible answers. Watson crashed several times during the second day of filming, nothing a regular viewer would notice while watching the recorded match on TV. One criticism I’ve heard about the match was that Watson was fed the questions electronically rather than relying on voice and character recognition. I have to agree that the electronic delivery could have been an advantage. Had voice recognition and OCR functionality been used in Watson, the victory would be quite a bit more impressive. I could clearly see the two mere mortals struggling to buzz in and shake in frustration when Watson was faster. The producers touted the physical buzzer plunger that Watson had to activate, but I still think that Watson had the advantage.
I would be interested to see a rematch in a year but with only inputs into Watson be voice and video of the Jeopardy board. After all, Deep Blue was given a second chance versus Garry Kasparov, so why not give the humans a second chance on more equal footing? It is quite possible that programming algorithms over the year would improve enough so that Watson still would win, despite the reliance of voice recognition and OCR. In that case, the victory would mean that much more. Even if you know the outcome already, I still recommend watching the matches. I already saw the Nova special, so I skipped through most of the background stuff from the IBM folks. Here is a link to my YouTube playlist that has the three episodes broken into 6 videos. Check it out! Then you can tell your grandchildren how you watched a computer beat humans in Jeopardy for the first time. Then they’ll ask, “Humans were allowed to play Jeopardy back then?”
Tonight with my wife, I watched an interesting episode of Nova titled “The World’s Smartest Machine.” Romantic, right? You can tell we are a couple of geeks at heart. Being a huge fan of Asimov and his robot creations, I am continually amazed at the progress of robotics and artificial intelligence. As a kid I was playing chess as early as the 3rd grade. I’ve always thought of myself as a pretty good chess player. When I heard about a computer that supposedly could beat a world champion chess player, I was intrigued. Garry Kasparov battled IBM’s Deep Blue supercomputer in 1996 and won, however he eventually lost a rematch in 1997. Well, IBM is back at it, this time taking on Jeopardy. Nova takes us behind the scenes with the developers of “Watson,” the next step in AI. I was disappointed when the episode ended and it didn’t say who won. I didn’t realize until logging on PBS.org that the contest actually was taking place while I was watching Nova (on my media center PC). After some frantic Google searches, I eventually came across a Twitter feed with a link to a Youtube posting of the episode. This was not quite 2 hours after the episode aired! Isn’t technology great? I expect the Youtube post to be served a DMCA take down, so watch it soon if you plan on it. In addition to the episode at the end of this post, I’ve embedded the Nova preview and an humorous parody of the challenges Watson overcame during his development. I suggest watching the Nova episode first. If you are at all interested in AI, you must check these out!
There are two more nights to the match, airing Tuesday February 15th, and Wednesday the 16th. Be sure to check them out, along with some live blogging on PBS.org. I find myself rooting for Watson, not because I want to see humans defeated, but because I want to see humans capable of making something smart enough to do it. Who do you think will win? Who do you want to win?
