作者位次 #2/4ACL 2026Poster· Natural Language Processing
KG - V i P : Bridging Knowledge Grounding and Visual Perception in Multi-modal LLM s for Visual Question Answering
Zhiyang Li · Ao Ke · Yukun Cao · Xike Xie
摘要
Multi-modal Large Language Models (MLLMs) for Visual Question Answering (VQA) often suffer from dual limitations: knowledge hallucination and insufficient fine-grained visual perception. Crucially, we identify that commonsense graphs and scene graphs provide precisely complementary solutions to t...