Wednesday, February 16, 2011

Jeopardy's Watson computer: Just a high-speed moron

By Donald Sensing

Here is the Youtube of IBM's computer, Watson, beating its human challengers like a mongrel dog on Jeopardy last night. Look carefully at the clues? Notice anything?



Almost none of the answers on the game board required any kind of abstract reasoning to answer. In fact, you could put a human with practically no knowledge of the subjects on the board in front of a computer connected to Google and that person could simply type in the nouns of the clues and get the same answers.

Example - the first question of the game:


And the Google:


Strip out Watson's blinding speed, and it is no smarter than human beings at all. Watson, for all its engineering impressiveness, simply did only what computers have always done: collate at blinding speed (and compute mathematical probabilities to choose an answer). It does not matter that Watson was not connected to the Internet since its mass-memory unit holds 16 Terabytes of data, processed by a 2,880 processor core. As my own computer professor said (many years ago!), "A computer is just a high-speed moron." There is nothing about Watson that I have read so far that obviates that observation.

I think its programmers must have realized this since they artificially crippled Watson by design. Note that Watson was programmed not to buzz in unless it computed an answer of at least 50 percent "confidence" of being right. This was an entirely artificial barrier. Why not 75 percent? Or 25 percent? Or any level at all? Watson is so blindingly fast that it could have buzzed in for every question before either champion (making their presence merely ceremonial, which it almost was anyway). Then Watson simply could have given its top answer, regardless of confidence level, and the computer still would have got 90% of the questions right.

In other words, if the game had allowed Watson to give uncrippled answers, it would have always answered first and would have won even more decisively. But then, with access to the same amounts of data resources through Google, I could beat the two champions if the game always allowed me to answer first. In fact, I'd win under the same confidence crippling as long as I could answer first - because clearly Watson will always be able to buzz in faster than a person.

So what does Watson really prove? From a technical, engineering and programming perspective, it's an amazing achievement with enormous potential for a wide range of applications ranging across broad multi-disciplinary subjects and problems. As for the Jeopardy game, there's less than meets the eye. Count it as a proof-of-concept exercise. What it did not do was reason abstractly. It just collated amazing amounts of information very rapidly. But we already knew that computers are faster than we are for specified tasks. That's why we build them to begin with.

Sorry, Prof. Reynolds, the Singularity has not arrived.

Update: Good discussion on this at The Speculist, including this nugget:
@stephentgo: Repercussions of IBM's Watson unknown, but any job that involves answering questions by phone will soon be at risk. http://bit.ly/hp91hW
Bookmark and Share