arXiv Computer Vision By Qixun Wang, Yang Shi, Letian Cheng, Zhuoran Zhang, Yan He, Yuqi Tang, Qi Zhang, Xinlei Yu, Ruizhe Chen, Tianrun Xu, Yuanxing Zhang, Pengfei Wan, Haotian Wang, Xianghua Ying

Beacon: Knowing When and How to Perform Agentic Visual Reasoning

Read the original on arXiv Computer Vision →

Beacon is a new agentic visual reasoning model that improves multimodal large language models (MLLMs) by better deciding when to use tools and how to use them. It introduces two key concepts—Mode Adaptiveness, which ensures tools are invoked only when necessary, and Tool Effect, which measures the net benefit of tool use— and trains the model with supervised fine‑tuning and reinforcement learning that rewards necessity-aware decisions and expands capability through expert hints. Across 13 benchmarks, Beacon outperforms other open‑source models, achieving the highest average score and the largest net tool‑gain on diagnostic tests.

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.

arXiv AI
Aug 3

Multimodal Reinforcement Learning with Adaptive Verifier for AI Agents

arXiv:2512. 03438v3 Announce Type: replace Abstract: Agentic reasoning models trained with multimodal reinforcement learning (MMRL) have become increasingly capable, yet they are almost universally optimized using sparse, outcome-based rewards computed based on the final answers.

By Reuben Tan, Baolin Peng, Zhengyuan Yang, Hao Cheng, Oier Mees, Theodore Zhao, Andrea Tupini, Isar Meijer, Qianhui Wu, Yuncong Yang, Lars Liden, Yu Gu, Sheng Zhang, Xiaodong Liu, Lijuan Wang, Marc Pollefeys, Yong Jae Lee, Jianfeng Gao
arXiv AI
6d ago

With Tools or Because of Tools? Tracing Capability Gains in Multimodal Agents

The paper examines whether tool-augmented multimodal agents truly gain capability from using tools. By comparing agents Thyme and DeepEyesV2 with tool-free and pure-text counterparts across tasks such as image understanding, OCR, chart reading, and math reasoning, the authors find that tool access offers little consistent improvement and that most problems solved with tools are also solvable without them. Ablation studies reveal that the full tool-use loop does not consistently outperform simpler tool-call or result-only formats, indicating that agents may learn tool-calling patterns rather than tool-contributed capabilities.

By Jiawei Guo, Donglei Yu, Yu Chen, Xiang Wang, Shuai Li, Xinpei Zhao, Huaxing Liu, Qinghao Wang, Minpeng Liao