AI RESEARCH

Target-Aligned Reinforcement Learning

arXiv CS.AI

ArXi:2603.29501v1 Announce Type: cross Many reinforcement learning algorithms rely on target networks - lagged copies of the online network - to stabilize