"->forside

Joel Lehman is an assistant professor at the IT University in Copenhagen, Denmark. He recieved his bachelor's degree from the Ohio State University in 2007, and his Ph.D. from the University of Central Florida in 2012.

He is co-author of the popular science book "Why Greatness Cannot be Planned: The Myth of the Objective," and is an inventor of the novelty search algorithm. Other research interests include neuroevolution, artificial life, and open-ended evolution.

Abstract:
One hallmark of natural evolution is its profound creativity, illustrated by the incredible diversity of complex life on Earth - ranging from bats to beetles to birch trees. If computers could be similarly creative, it would be possible to autonomously generate a diverse complexity of useful robots.

This talk describes creative evolutionary algorithms, inspired by natural evolution, that can be used to evolve simulated brains for robots. The main idea is to breed robots to behave in continuously novel ways, based on the insight that a new robot behavior may be the gateway to further novelty. This technique is demonstrated in a variety of domains, from wheeled robots navigating a maze, to teaching a biped robot how to walk, to evolving both the morphology and brain of locomoting 3D virtual creatures. Interestingly, breeding robots towards novelty often results in better-functioning robots than does directed breeding towards ever-increasing performance. The talk concludes with reflections about what this strange result seems to suggest about natural evolution and more broadly, about achieving ambitious objectives in general.

Link:
Joel Lehman

[image source: Joel Lehman]
[source: Vive Les Robots!]