It’s been barely three years since we introduced our neural network add-in to Excel. It wasn’t the first decision analysis tool of its kind, but at the time, neural network technology was talked about only in futuristic terms–potential applications. Now, as I write this, these applications-to-be–as well as some we never speculated on–actually are. Neural nets have been used for everything from wool pricing in Australia to teaching a robot how to walk in rough terrain. This trend has just been romanticized by a Canadian nerd-turned rapper in her homage to artificial intelligence.
The disciplines where neural network technology has moved most quickly to the forefront are medicine research and robotics development. Both fields use the human body as a model, and, to cite evidence of how far neural network technology has come, a group of National Science Foundation grants announced earlier this month include one for “understanding how the brain visually recognizes objects.” This project will use a computer analog of the brain to understand the brain itself.
Anywhere that enough historical, or even theoretical, data can be summoned, a neural network can make sense of it. Which makes me wonder why Wall Street–where decision making under uncertainty and risk assessment have just become contact sports–hasn’t made more extensive use of the technique. Or maybe the financial jocks are, in fact, using this form of classification and prediction. Anybody out there know of financiers who do neural nets?