Variance-Averse $n$-Step Offline Reinforcement Learning for Sparse Long-Horizon Environments

Excerpt

arXiv:2610.07899v1 Announce Type: new Abstract: Generative actors are transforming offline reinforcement learning (RL) by enabling expressive policy classes that model complex action distributions. However, this expressiveness also exposes a key challenge in heterogeneous datasets: generative policies can reproduce unreliable action modes whose return distributions exhibit high variance, occasionally yielding high returns by chance but lacking consistency. Consequently, maximizing the expected $