Analyzing a Decade of Human-Competitive (“HUMIE”) Winners: What Can We Learn?

Abstract

Techniques in evolutionary computation (EC) haveimproved significantly over the years, leading to asubstantial increase in the complexity of problems thatcan be solved by EC-based approaches. The HUMIES awardsat the Genetic and Evolutionary Computation Conferenceare designed to recognise work that has not just solvedsome problem via techniques from evolutionarycomputation, but has produced a solution that isdemonstrably human-competitive. In this chapter, wetake a look across the winners of the past 10 years ofthe HUMIES awards, and analyse them to determinewhether there are specific approaches that consistentlyshow up in the HUMIE winners. We believe that thisanalysis may lead to interesting insights regardingprospects and strategies for producing further humancompetitive results.