Ryan’s Blog Just another WordPress weblog

IBM and Watson

It's been two years since I've even thought about using this blog. However, I feel strongly enough about this subject, and will talk about in depth enough that I think this is the only logical outlet.

When I told my wife about Watson, she couldn't understand why I was so excited. She said, "It's a computer. It knows everything, so of course it's going to win." It took me nearly an hour to explain the gravity of the situation, and I think she finally decided to see my point of view. I hope I can do the same here for other folks.

Computers are amazing to me. They can do some of the most amazing things, and sometimes, I find myself shaking my head and saying, “Wow, that's cool” or “Unbelievable.” Even some of the simplest things, like a word processor, are enough to make me appreciate what's really happening. A word processor can make what used to be yesterday’s major task, and turn it into some of today’s most mundane and unthoughtful of processes. For example, I've got to make a flier for the bake sale. No big deal right? Fire up Open Office, type in who, what, when and where, paste in a picture or two, change a font here and there, add a couple colors, print, bam, you're done. Literally, that fast, done in just minutes. How long would it take to make a flier without one?

Think about that. You've got to get the paper, grab a pencil, sketch out what you want. Maybe cut out a picture or two from a magazine, and tape them on. Want some color? Grab some pens or markers and highlight here and there. Oops, I got the time wrong. I have to start over. Oh, I'd rather use blue than red, I gotta start over. Just making a simple change would take exponentially longer than the one or two mouse clicks we are used to.

Now, think about the Internet. In its most barebones, elemental, simplistic form, what's going on there? A computer takes a little blip of electricity, puts it on a wire and sends it to another computer. That second computer picks it up, figures out what it is, does something with it, and sends another blip across that wire back to the first computer, which does the same. Oh, and those computers are hundreds of miles apart. Little tiny blips of electricity are being sent across wires, between hundreds of computers just to get from one side of the country to another. And, all of these computers can do something with that electricity, process it, and turn it back into something that's meaningful to us. How long does it take for that to happen. Microseconds. That's incredible!

Computers and technology have, and continue, to do amazing things. 100 years ago, if you wanted to send a message from one side of the country to another, it would mean someone would have to physically take that message back and forth. It would be weeks, or months for that to happen. Sometimes, and entire team of people would take on the task. And if the task was even successful, not everyone would have survived the attempt. 50 years ago, it might have taken a few days. 20 years ago, phone calls were common, and you could chat with anyone almost any time. Now, if John in New York wants to talk to Dave in California, he pulls his iPhone out of his pocket, punches a few numbers, and John and Dave are having a face to face video chat, instantly. We have shrunk the world.

In the 1980's, it was thought that no computer could ever beat a human opponent in a game of chess. In 1997, IBM’s Deep Blue bested Garry Kasparov in a six game series. IBM had done the impossible, and created a computer that defeated, not just a human opponent, but reigning Chess Grandmaster, multi-time world champion, who many people to this day, still believe to be the greatest player of all time.

In the game of chess, there is a finite number of possibilities. The computer was able to analyze every possible move, and chose the one that was most likely to lead to the end result of winning the game. Computers are great at this type of concrete thinking. They can sort, organize and compute faster than imaginable. Where they fail is in comprehension.

Human language is one of the most complicated, fluid, and open ended concepts. Puns, word play, sarcasm, slang, accents, emphasis, the list of modifiers and variables is almost endless. Computers have a very difficult time interpreting language because the rules are so open. Nothing is concrete, and computers aren't so good at that.

What did IBM decide to do? The looked at what computers can do well, and not so well. They wanted to take on the challenge of human language. They wanted to create a computer that could basically master language, interpret what was wanted, and present the answer.

What competition requires a master of human language? The game show, Jeopardy! You will find everything but a straight question on that show. Sometimes, even I don't even know what's really being asked for. The show is full of jokes, word play, subtleties, and all sorts of complex language. IBM decided they wanted a computer that could not only comprehend the questions, put present the answer as well.

IBM's Watson is tackling the biggest hurdle computers have faced since their inception. And so far, it appears to be succeeding.