SUNY New Paltz Professor Strikes Gold!
Techniques of genetic and evolutionary computation are being increasingly applied to difficult real-world problems—often yielding results that are not merely academically interesting, but competitive with the work done by creative and inventive humans. These biologically motivated techniques begin with random data and grow or evolve solutions to problems that were never imagined by their human designers.
This month the ACM Special Interest Group on Genetic and Evolutionary Computation held their annual Genetic and Evolutionary Computation Conference (GECCO) in Atlanta, Georgia. For the past five years a central theme of these conferences has been the competition for Human-Competitive awards. This year \$10,000 in prizes went to contestants who used genetic and evolutionary algorithms to automate discoveries that would have been viewed as competitive achievements had the same discoveries been make by humans.
The $5,000 first prize in the 2008 Human-Competitive competition went to David Clark (Mathematics, SUNY New Paltz), Lee Spector (Cognitive Science, Hampshire College) and three Hampshire College undergraduates: Bradford Barr, Jon Klein and Ian Lindsay. In their paper, "Genetic Programming for Finite Algebras", the team reported discoveries obtained by genetic programming that they demonstrated to improve past human and computer results by up to 14 orders of magnitude.
"This year's field had a number of exciting and impressive entries," said Clark. "Ours stood out because of the long documented history of efforts to solve these problems by both theoretical and computational techniques. We were able to give a quantitative argument that our results were not only human-competitive, but in fact exceeded past human efforts by a substantial margin."