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