English(EN) PerceptRL: Making Visual Perception Rewardable in MLLMs from Answers to Interactions

PerceptRL框架通过代理强化学习提升MLLM视觉感知能力

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-28 04:00

研究人员开发了PerceptRL，一个旨在增强多模态大语言模型（MLLM）视觉感知能力的新框架。该系统通过使用多代理管道提取图像标题中经过验证的错误，并将其转换为带有基于规则的奖励的问答对，从而解决了细粒度视觉感知方面的局限性。PerceptRL还结合了多轮检查和视觉工具来处理复杂的感知问题，并纠正空间操作以提高工具使用场景的准确性。在七个基准测试的评估中，PerceptRL的平均准确率为71.6%，显著优于基础模型和前沿的视觉-语言模型。 AI

影响增强了MLLM的视觉感知能力，有望提高在需要详细图像理解和工具使用任务中的性能。

排序理由该集群描述了一篇关于改进MLLM感知的新型框架的最新研究论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 12 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群描述了一篇关于改进MLLM感知的新型框架的最新研究论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Tongkun Guan, Zhibo Yang, Mingkun Yang, Haocheng Wang, Wei Shen, Xiaokang Yang · 2026-10-28 04:00

PerceptRL：使视觉感知在多模态大模型中从答案到交互可奖励

arXiv:2610.37471v1 Announce Type: cross Abstract: Fine-grained visual perception remains a bottleneck during post-training of multimodal large language models (MLLMs), as perceptual errors can constrain gains from reasoning-focused optimization. Although image-caption pre-trainin…

本文如何被排名

报道来源 [1]

PerceptRL：使视觉感知在多模态大模型中从答案到交互可奖励

相关实体

相关话题