Okay, so this blog has nothing to do with Sherlock Holmes or Dr. Watson, but if I were to title the blog post “Thomas J. Watson’s Super Computer takes on Jeopardy Champs” I would lose the effect of the analogy. And besides, I don’t really care for search engine rankings and I love to write for the sake of writing. So fans beware, I’m about to trash talk Dr. Watson’s good name.
In a few days, Watson, the IBM Super Computer will take on two of the most celebrated Jeopardy Champs, Ken Jennings and Brad Rutter. I happen to have studied a bit of artificial intelligence and can reasonably tell a story of why Dr. Watson is always doomed to follow Sherlock Holmes’ lead and why should that matter to Alex Trebek. Bear with me.
In the beginning, Dr. Watson became Sherlock Holmes’ sidekick, following every cue and provided no initiatives deserving to be mentioned. Hence the Hollywood made movies about Sherlock Holmes and not Dr. Watson. Then there are computers, following our every command and never making a suggestion to us on how to be human. But the dark side of the force wishes to bring upon us the wrath of Matrix. They want to create machine intelligence to rival our own and challenge our existence.
Rewind to 2011, phew… it’s a good thing we are not there yet.
Machine intelligence is not new and it is definitely evolving. Remember Deep Blue - the supercomputer that beat world chess-champ Kasparov? The way the machine operated then was by sheer calculation of the possible moves in binary and react to Kasparov’s mistakes. The machine itself had no intelligence, mere algorisms and a databank. Then came “Eliza,” who followed simple rules of grammar and syntax and responded to human input with the appearance of a conversation. For example, if you were to ask Eliza “how do you feel today” Eliza would respond with something like: “does that question interest you?” Eliza obviously had not understanding of the human condition and merely reacted logically to a conversation. Even though Eliza had logical operations, she did not have an advanced logical routine. She merely followed a pre-defined script without maxims. Then came along a computer that had general logical reasoning. I can’t remember what our nefarious scientists called it, but it too lacked an understanding of the basic human rules of conversation.
To challenge Jeopardy, IBM had to create a computer that can learn, can process pattern recognition. Not just the simply kind of pattern recognition, spotting a circular shape to match with the letter “O.” Watson had to have the kind of complex pattern recognition that can handle Jeopardy’s unique experiment with human understanding. Watson will also need enormous memory storage since it would not be plugged into the Internet.
Watson is no chump-change. It is the equivalent of 6,000 high-end home computers and has a storage containing over 10 million documents. But it is still about pattern recognition and calculation based on best matches of possible results. It is still no Sherlock.
Here is the moral of the story. Computers are finding patterns and learning about us, but it will always just react to our needs. The few industry market geniuses are analyzing your patterns using these supercomputers and influencing you in irresponsible ways that destroyed your relationship with food and destroyed our planet.In our story, Dr. Watson is helpless to us because it cannot engage our needs in the conversation. No matter ho hard it tries, it cannot fight our fight for us. Sherlock must do the work. It's time we take back that responsibility and take the initiatives, being Sherlocks, and lead ourselves in this technology driven world. No offense to Watson - the computer, but until we have control back from the industries that are exploiting our trust and your good will, we have to hope Ken Jennings kicks your ass.