Frictional Q-Learning algorithm enhances reinforcement learning stability and performance

Researchers have introduced Frictional Q-Learning, a novel off-policy reinforcement learning algorithm designed to address extrapolation errors. By drawing an analogy to static friction, the method models the replay buffer as a low-dimensional manifold and identifies supported actions as tangent directions. This approach encodes supported actions using a contrastive variational autoencoder, leading to more stable and robust performance on continuous-control benchmarks compared to existing methods. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Introduces a novel method to improve stability and robustness in off-policy reinforcement learning, potentially enhancing performance in complex control tasks.

RANK_REASON This is a research paper detailing a new algorithm for reinforcement learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Hyunwoo Kim
• arXiv
• Frictional Q-Learning

• paper
• other