AI RESEARCH
iReasoner: Trajectory-Aware Intrinsic Reasoning Supervision for Self-Evolving Large Multimodal Models
arXiv CS.CL
•
ArXi:2601.05877v3 Announce Type: replace Recent work shows that large multimodal models (LMMs) can self-improve from unlabeled data via self-play and intrinsic feedback. Yet existing self-evolving frameworks mainly reward final outcomes, leaving intermediate reasoning weakly constrained despite its importance for visually grounded decision making. We propose iReasoner, a self-evolving framework that improves an LMM's implicit reasoning by explicitly eliciting chain-of-thought (CoT) and rewarding its internal agreement.