← Back to all news
arXiv Computer Vision By Zixiu Ding, Zilin Zhao, Yingjie He, Xinlang Kang, Guansu Wang, Wei Zhang

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

Read the original on arXiv Computer Vision →

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.

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 Computer Vision.

Related stories

arXiv AI
Aug 12

Grounded Post-Training with Hard Examples for Reducing Hallucination in Multimodal Large Language Models

arXiv:2605. 16411v3 Announce Type: replace-cross Abstract: Hallucination remains a fundamental challenge in vision-language models (VLMs), where autoregressive generation may produce linguistically plausible yet physically inconsistent or visually ungrounded responses due to likelihood maximization under joint probabilistic modeling.

By Qinwu Xu
llmscomputer-visionreinforcement-learningmultimodalbenchmarkssafety
More like this →
arXiv Computer Vision
Oct 23

Visual Grounding First, Multimodal In-context Learning Follows

The paper introduces ProCoRe, a contrastive reinforcement learning approach that improves visual grounding in multimodal large language models (MLLMs) for in-context learning (ICL). By generating captions for contrastive image pairs, decomposing them into propositions, and optimizing contrastive rewards, ProCoRe enhances alignment with relevant visual evidence without requiring ICL-formatted training data. The method achieves significant gains in multimodal ICL classification accuracy and captioning ROUGE-L over the Qwen-VL-3-8B baseline and outperforms state‑of‑the‑art ICL‑data‑dependent methods across multiple benchmarks. Additionally, the authors present PICL, a personalized multimodal ICL captioning benchmark designed to test genuine understanding of visual demonstrations.

By Minhyuk Seo, Minjae Lee, Chaeeun Lee, Wei Lin, Muhammad Jehanzeb Mirza, Tinne Tuytelaars, Jonghyun Choi
llmsreinforcement-learningmultimodalbenchmarkssafety
More like this →
arXiv Machine Learning
Jun 26

Staying VIGILant: Mitigating Visual Laziness via Counterfactual Visual Alignment in MLLMs

arXiv:2606. 26387v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) extend large language models (LLMs) with visual perception, enabling joint reasoning over images and text.

By Xi Xiao, Chen Liu, Chih-Ting Liao, Yunbei Zhang, Qizhen Lan, Yuxiang Wei, Lin Zhao, Janet Wang, Jianyang Gu, Muchao Ye, Tianyang Wang, Hao Xu
llmsreinforcement-learningmultimodalbenchmarkssafety
More like this →