一作ACL 2026Findings· Findings of ACL
See or Say Graphs: Agent-Driven Scalable Graph Understanding with Vision-Language Models
Shuo Han · Yukun Cao · Zezhong Ding · Zengyi Gao · S. Kevin Zhou · Xike Xie
摘要
Vision-language models (VLMs) have shown promise in graph understanding, but remain limited by input-token constraints, facing scalability bottlenecks and lacking effective mechanisms to coordinate textual and visual modalities. To address these challenges, we propose GraphVista, a unified framew...