Fast Value Tracking for Deep Reinforcement Learning: Leveraging Kalman Filtering Paradigm

Efficiently draw samples from posterior distribution for robust reinforcement learning.