Dean Barr On Artificial Intelligence by Thom Hartle
Barr explains the differences between expert systems and neural networks and some of the keys to successfully applying neural network technology to trading. He also looks at the future of neural network technology.
How did you get started managing money?
I started on Wall Street in 1983 in a Merrill Lynch training program. Prior to that, I was fortunate enough
to spend a few summers on a trading desk. When I joined Merrill Lynch in 1983, I was shipped off to
Europe for about a year, where I learned a great deal about the international markets and the way they
operate, particularly on the equities side.
And after working for Merrill Lynch?
I was recruited by Goldman Sachs and returned to New York. There I was exposed to emerging trading
and analytical technologies, both in portfolio restructuring as well as trading, through some of my larger
institutional clients. I was working with a lot of large hedge funds, arbitrageurs and big index-type
Was your educational background mathematical?
Actually, it was a smattering of finance and computer science. I did my undergraduate work at Cornell University and received an MBA from New York University. I was fortunate to be exposed to a lot of very
bright people. Many new creative, innovative techniques were evolving. The thing that really struck me
— which backs into where I am today — is that there are a lot of ways to make money in the markets. A
lot of my clients were making real money in the markets. I was exposed to the good, the bad and the ugly
in money management: firms that did well, those that did okay and those that did poorly.
What kind of ideas did you come away with?
I felt that inherent in the market environment the opportunity to make money was there, but you had to
have an edge. I thought that the edge could be technology-driven — that's where I thought the future