Reinforcement-Learning
3 posts
A Cooperative Benchmark: Announcing the Hanabi Learning Environment
Announcing the Hanabi Learning Environment, a research platform for multiagent learning and emergent communication.
Classic and Modern Reinforcement Learning
What is deep reinforcement learning, really? A reflection on the field and the motivation for this blog.
Introducing the ALE 0.6
ALE 0.6 adds modes, difficulties, and sticky actions — a richer framework for testing AI agent generalization.