Evolutionary Intelligence Seminar (Online) - Risto Miikkulainen - Neuroevolution
It is my pleasure to announce that, for the occasion of the upcoming PhD defense of Arkadiy Dushatskiy on September 6 (https://www.tudelft.nl/en/events/2023/tu-delft/thesis-defense-a-dushatskiy and https://www.cwi.nl/en/news/evolutionary-algorithms-for-optimization-and-medical-image-segmentation/) Prof.dr. Risto Miikkulainen from the University of Texas at Austin will visit CWI and give a talk on neuroevolution. Prof. Miikkulainen (https://www.cs.utexas.edu/users/risto/) is a Professor of Computer Science at the University of Texas at Austin and AVP of Evolutionary Intelligence at Cognizant AI Labs. He received an M.S. in Engineering from the Helsinki University of Technology (now Aalto University) in 1986, and a Ph.D. in Computer Science from UCLA in 1990. His current research focuses on methods and applications of neuroevolution, as well as neural network models of natural language processing and vision; he is an author of over 450 articles in these research areas.
You are welcome to join. As a conference is being hosted at CWI at the same time, the room that the talk is held in unfortunately has limited seating capacity. We will also stream the talk online through Zoom. Details can be found below.
Best,
Peter
Date: Tuesday 5 September 2023, 16:00
Location: CWI, room M290
Zoom: https://cwi-nl.zoom.us/j/81815944835?pwd=K24zMSs3bmFYMEcvMHRlQzF3T0NqZz09
Title: Leveraging human expertise with machine discovery
Abstract:
Given a surrogate model of the world, neuroevolution can discover effective decision strategies that human decision makers may miss. However, humans may also have expertise that can be instrumental in constructing good strategies. If such expertise is distilled into a canonical neural network representation and used as an initial population, the evolved solutions may exceed both human solutions and those evolved from scratch. Further, evolution can discover how to utilize the hidden potential of good ideas in otherwise weak solutions. I will demonstrate these principles in a synthetic domain as well as in the real-world domain of discovering non-pharmaceutical intervention policies in COVID-19.
