Emmanuel Rachelson

A new framework improves optimization in machine learning using behavior information.

Introducing ExpectRL to tackle overestimation in Reinforcement Learning through expectiles.

A new benchmark for testing robust reinforcement learning methods in various environments.

Researchers enhance reinforcement learning with a new framework for uncertain environments.

LEADS method enhances skill diversity for better AI exploration.

SGDrop helps CNNs learn better from limited data by broadening their focus.

RAMP helps agents explore new paths for better learning.