作者位次 #5/7ACL 2026Poster· Natural Language Processing
Simple- VGC : Enhancing Visual Grounding in Multimodal Reasoning via Adaptive Tool Composition
Ye Wang · Qianglong Chen · Siyuan Wang · Zejun Li · Shijie Guo · Zhirui Zhang · zhongyu wei
摘要
Multimodal Large Language Models (MLLMs) have achieved strong performance on vision-language tasks, yet often fail to preserve and effectively leverage visual evidence throughout generation. We identify three fundamental types of visual grounding failures: Long-Context Grounding Error, where visu...