RL-CPFusion

Offline Reinforcement Learning with Conformal Prediction

Tackle real-world decision-making challenges with our innovative approach that combines the power of Offline Reinforcement Learning and the reliability of Conformal Prediction.

Offline Learning

Learn effective policies from static datasets without direct environment interaction.

Uncertainty Quantification

Leverage Conformal Prediction for robust confidence intervals on Q-value estimates.

Stable & Reliable

Mitigate overestimation risks and ensure a more stable learning process.