4 papers
Exploration by Running Away from the Past
Paul-Antoine Le Tolguenec, Yann Besse, Florent Teichteil-Koenigsbuch +2
The ability to explore efficiently and effectively is a central challenge of reinforcement learning. In this work, we consider exploration through the lens of information theory. S…
Exploration by Learning Diverse Skills through Successor State Measures
Paul-Antoine Le Tolguenec, Yann Besse, Florent Teichteil-Konigsbuch +2
The ability to perform different skills can encourage agents to explore. In this work, we aim to construct a set of diverse skills which uniformly cover the state space. We propose…
Genetic Drift Regularization: on preventing Actor Injection from breaking Evolution Strategies
Paul Templier, Emmanuel Rachelson, Antoine Cully +1
Evolutionary Algorithms (EA) have been successfully used for the optimization of neural networks for policy search, but they still remain sample inefficient and underperforming in…
Quality with Just Enough Diversity in Evolutionary Policy Search
Paul Templier, Luca Grillotti, Emmanuel Rachelson +2
Evolution Strategies (ES) are effective gradient-free optimization methods that can be competitive with gradient-based approaches for policy search. ES only rely on the total episo…