← Back to all news
arXiv Machine Learning By Haocun Ye, Xinlong Jiang, Qile Chen, Bingyu Wang, Teng Zhang, Shubai Chen, Tingyu Wu, Zhenkun Zheng, Yiqiang Chen

Same Reward, Different Skills: When Multimodal RL Learns to Look

Read the original on arXiv Machine Learning →

The paper demonstrates that reinforcement learning with verifiable rewards (RLVR) can improve vision‑language benchmark performance even when models are trained without visual input. When real images are introduced at test time, models trained blind recover about half of the performance gain at 3B parameters and nearly four‑fifths at 7B, but extended real‑image training can erode grounding while benchmark gains persist. The authors propose a visual resolvability rule and show that requiring visual evidence for correct answers leads to significant improvements in target discovery and generalization to unseen question types, while controls confirm that the gains stem from actual visual grounding rather than artifacts.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv Computation and Language
Sep 4

Attend to Evidence: Evidence-Anchored Spatial Attention Supervision for Multimodal RLVR

The paper introduces EASE, a method that enhances multimodal reinforcement learning with verifiable rewards (RLVR) by adding visual‑evidence process supervision. EASE transforms annotated evidence regions into smoothed visual‑token targets and uses them to guide attention during RL training, but only on high‑reward trajectories. Experiments on Qwen2.5‑VL‑7B, Qwen3‑VL‑4B, and Qwen3‑VL‑8B show that EASE improves average scores over DAPO by 2.5 to 3.1 points across perception, hallucination, visual math, and multimodal reasoning benchmarks, and diagnostics confirm better alignment of visual attention with annotated evidence.

By Ruina Hu, Chen Wang, Lai Wei, Jionghao Bai, Bin Yu, Weiran Huang, Kai Wang, Yue Wang
llmsreinforcement-learningmultimodalbenchmarkssafety
More like this →
arXiv Computer Vision
Sep 16

SAVOR: Self-Aware Visual Grounding via Confidence-Calibrated Reinforcement Learning for Multimodal Hallucination Mitigation

SAVOR is a training framework for multimodal large language models that adds token and answer confidence to the output schema, optimises a Group Relative Policy Optimisation objective to penalise calibration error and poor abstention, and uses the learned confidence at inference to revisit visual evidence only when uncertain. Experiments on POPE, HallusionBench, AMBER, and MMHal-Bench with InternVL3-8B and Qwen3-VL-8B backbones show that SAVOR reduces hallucination while maintaining general capability on MME and MMBench, achieving lower Expected Calibration Error than DPO and decoding baselines.

By Zixiu Ding, Zilin Zhao, Yingjie He, Xinlang Kang, Guansu Wang, Wei Zhang
llmsnlpreinforcement-learningmultimodalsafety
More like this →