What Ever Happened to Artificial Intelligence?

Artificial intelligence (AI) is the intelligence of machines and the branch of computer science that aims to create it. AI textbooks define the field as "the study and design of intelligent agents" where an intelligent agent is a system that perceives its environment and takes actions that maximize its chances of success. - Wikipedia

ar·ti·fi·cial meaning humanly contrived often on a natural model : man-made such as an artificial limb or artificial diamonds.  in·tel·li·gence meaning the ability to learn or understand or to deal with new or trying situations.  So were are these man-made things that can understand their environment and learn from their mistakes?  Artificial Intelligence (AI) has been talked about since the early days of computing.  Sometimes it has been in the forefront of our minds, but mostly it is discussed in the dark hidden away labs of the worlds universities and larger computing companies.  The stuff really of science fiction, AI doesn't really resinate with the common person because there is nothing for them to really see, until this past week.

Courtesy of IBM Research

But first we need to go back some 14 years to 1997, when I was first introduced to AI. In May 1997, IBM's Deep Blue Supercomputer played a chess match with the reigning World Chess Champion, Garry Kasparov (this was actually the second time there was a smaller match a year earlier where Kasparov won). The event was captured live and broadcast over the web, where millions of chess and computing fans tuned in to witness the event in real-time.  I had the opportunity to attend the event in NYC.  Now I know how to play chess (mostly loose at chess), but this was the first time I saw a real chess match. They played six games over six days.  I remember the match I saw where Kasparov was getting really mad because he was loosing. At the end of the sixth day the Deep Blue 2 Kasparov 1 Draw 3. I am not sure if this was a testimony to Kasparov or the fact that Deep Blue learned how to tie and ultimately defeat him. A lot of effort went into this but the scientific and technology community learned a lot and Kasparov pouted for awhile but he continued to play computers to gain his revenge.  He played computers publicly twice more in January 2003 against Deep Junior where he offered a draw and again in November 2003 when he played against a computer wearing 3D glasses, which also ended in a draw.  Even though Kasparov insisted that the computers were some how cheating, what was happening was they were learning from a master.


Now lets fast forward 14 years to this past week.  The big news (hype) of the week for AI fans has been the Jeopardy match between IBM's Watson computer and the two all time best Jeopardy players; Ken Jennings and Brad Rutter.  Now the point of this exercise is to demonstrate that a computer can go through a huge database of information in a lighting fast way, learn form its mistakes and present answer to common questions and do this better than a human being.  This is really not that big a deal.  The big break through here is that it does this naturally (technical term is that Watson uses algorithms to understand the natural language that information is written in) in other words Watson can understand the different meaning of the word bat for example.  Watson knows the difference between a baseball bat and the animal bat and unlike its predecessors Watson can hear the question, see the environment and speak the answer naturally.  Deep Blue and the others couldn't do this because the technology was not available, but not today.  This is highly complex because Watson has to "understand" the questions.  Here's how Dr. Jon Lenchner of IBM Research explains it:

“The clue was: It was introduced by the Coca-Cola Company in 1963. Watson can find a passage stating that ‘Coca-Cola first manufactured Tab (the correct response) in 1963’, so in order to answer the question, Watson needed to understand that introducing and manufacturing can be equivalent – if a company is introducing a product. But that is highly dependent on context: if you introduce your uncle, it doesn't mean you manufactured him.”

Human's understand this with ease, but computers don't and that's where AI comes in. Human's also can do this with a device that weighs about 3 pounds and takes up very little space; the brain.  Watson on the other hand requires custom algorithms, terabytes of storage and thousands of POWER7 computing cores working in a massively parallel system in order to answer the questions in the same time as a human.  So for today this wonder of AI is big and not very portable.  But, so was the first computer the ENIAC, short for Electronic Numerical Integrator And Computer, also known as the "Giant Brain." It was a large calculator used by the military in World War II to calculate artillery firing tables.  The point is that eventually large computing things get smaller.  So it's only natural that the computing power of Watson will someday be the same size let's say of the human brain. But that's the stuff of science fiction, or is it?


Until the technology to reduce Watson in size becomes available (and it will) there are a lot of applications for AI that we should be moving forward with.  It's already being used by companies like Google to improve the search experience, but this is only the tip of the iceberg.   Complex problems that are taking thousands of human beings to solve could be done by one or two Watson (think cures for disease) and jobs that require hundreds of humans doing complex operations (air traffic controllers for example) could be replaced by a few Watson's. Could AI be used to interrupt facial expression and body movements during airport security screens and parse out the bad guys, absolutely. Now I know what your thinking you Terminator fans, but lets not cloud our judgement based on some 1984 science fiction movie.  AI is finally getting to the point of some real world applications, but before it can move forward human's might have to take a step to the left and put a little faith in these computers.


Now that Watson has successfully "won" at Jeopardy what's next for the super computer?  Watson is moving to NYC to the Columbia University Medical Center where it will be used to help doctors evaluate patients at the facility, helping to reduce the time needed to evaluate and determine the correct diagnosis for a patient.  Watson will also lend doctors a hand when it comes to developing more "personalized treatment options" for patients' needs through use of the Watson software.


Interested in learning more about IBM's Watson. Watch these videos.

    

Watch the full episode. See more NOVA.

Contributions
IBM Research at http://www.research.ibm.com/
Watson Research Center at http://watson.ibm.com
DNA info.com
PBS at http://www.pbs.org/wgbh/nova/tech/smartest-machine-on-earth.html