Author - whirl

Three ICML papers from WhiRL members.

WhiRL members have contributed three papers to this year’s ICML: Stabilising Experience Replay for Deep Multi−Agent Reinforcement Learning Programming with a Differentiable Forth Interpreter Input Switched Affine Networks: An RNN Architecture Designed for Interpretability  

ICML paper accepted

Our paper Stabilising Experience Replay for Deep Multi-Agent Reinforcement Learning was accepted for publication at ICML-17.