I am a postdoc at Laboratoire d'Informatique de l'École Polytechnique (LIX) with Johannes Lutzeyer and Jesse Read. We work on reinforcement learning and graphs. During my PhD, I studied interpretable reinforcement learning at Inria Scool in Lille. I designed algorithms and benchmarks to make RL agents verifiable, transparent, and reproducible.
Hector Kohler
hector dot kohler at polytechnique at edu
Scholar
Github
Palaiseau, France
Teaching assistant
-
Decision under Uncertainty and Reinforcement Learning (Grads, 25h)
Sequential Decision Making (Grads, 6h)
Algorithms and Programming II (Undergrads, 25h)
Algorithms and Data Structures (Undergrads, 36h)
Notebooks
Supervisions
-
Image Segmentation (1 month)
Entropy Regularized RL (2 months)
Object-Centric Atari RL (6 months)
Policy Distillation to Trees (6 months)