Kenneth Stanley will present his work on Novelty Search and open-ended evolution at Fabelier, Monday 21th of May, 7pm.
Video streaming by Ustream
Ken is the director Evolutionnary Complexity Research Group at University of Central Florida.
This will be a mind-blowing talk ! #Evolution #MachineLearning #AI #Novelty #Research #Discovery
I’ll try to put some streaming on Fabelier’s USTREAM channel.
While search and optimization algorithms almost always aim for a specific objective or set of objectives (including evolutionary algorithms), in this talk I will highlight several reasons to believe that search works best without explicit objectives. The talk will explore the benefits of non-objective search, including a new search algorithm called novelty search, as well as several hidden costs of explicit objectives that are rarely discussed. By recognizing the limitations of traditional approaches to search, the hope is that we can liberate automated search algorithms (such as evolutionary algorithms) to discover artifacts that were heretofore out of reach.
WHEN : http://fabelier.org/events-agenda/
See you there !