Preprint
Aug 2026
Q-CueGraph: Query-Conditioned Visual Evidence Graphs for Multimodal Reasoning
Q-CueGraph maps a question and an image representation to budgeted, coordinate-level observations for a frozen reader, and reaches 92% of full-image ANLS on InfographicVQA from about half the image area.
Peng Pan, Xinfang Zhang
· 0 citations