Reinforcement Learning with Visually Attributable On-Policy Self-Distillation

Paper recorded by Signals 4 on 2026-10-19 in cs.AI. Abstract reproduced from arXiv; link to the original below.

Published 2026-10-19 on arXiv · recorded by Signals 4 on 2026-10-20

Category: cs.AI · 人工智能 · first seen 2026-10-20

Abstract

Multimodal reinforcement learning with verifiable rewards (RLVR) provides reliable task-level supervision, yet offers limited guidance for localizing errors in visual reasoning. On-policy self-distillation (OPSD) complements this supervision with dense token-level corrections from a privileged teacher equipped with an evidence-centered visual view. However, the resulting corrections may conflate v

Read on arXiv →