ConceptDRL / Distributional Reinforcement Learning

Distributional Reinforcement Learning

A subfield of reinforcement learning that learns the full probability distribution of returns rather than only their expectation, thereby capturing uncertainty and supporting risk-sensitive decision making.

Timeline (last 90 days)

Oct 22

Distributional reinforcement learning (DRL) learns the full distribution of returns rather than only their expectation, an idea first systematized by Bellemare et al. in the C51 algorithm.

Unverified50%

All Facts (1)

Unverified

Distributional reinforcement learning (DRL) learns the full distribution of returns rather than only their expectation, an idea first systematized by Bellemare et al. in the C51 algorithm.

50%

Source Articles

Quantile Distributional Reinforcement Learning Achieves Near Minimax Rates: Theoretical Breakthrough and Statistical Accuracy AnalysisOct 22