On April 22nd, Paul Templier successfully defended his PhD, entitled Leveraging Structures in Evolutionary Neural Policy Search (abstract below). Paul’s thesis was the first that I officially co-directed, as opposed to co-advised, and the thesis funding, from ISAE-Supaero and the Région Occitanie, was one of the first grants I received. Paul advanced the field of evolutionary neural policy search with multiple contributions on neural network representation, Quality Diversity, and combining evolutionary and RL methods. Paul will be continuing his research in the lab of Antoine Cully, so I’m really excited to see what great research he does in the future.
Leveraging Structures in Evolutionary Neural Policy Search
While training an artificial agent for complex tasks like driving a car, mastering a video game, or controlling plasma in a nuclear fusion reactor, innovations can lead to intelligent behavior. In such scenarios, a promising approach is to mimick the natural world’s evolutionary process, which has honed the problem-solving capabilities of animal brains. Evolutionary Neural Policy Search (ENPS) draws inspiration from this concept. It creates a diverse population of “brains” represented by neural networks, allowing the system to “evolve” by selectively combining and mutating successful individuals. This thesis delves into the core components of ENPS and their intricate interplay. By analyzing the structures of ENPS, the goal is to design novel policy search methods that enhance these components, ultimately leading to the development of more efficient and effective learning algorithms for complex tasks.