AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Uncovering Grounding IDs: How External Cues Shape Multi-Modal Binding

Amirmohammad Izadi, Hosein Hasani, Fatemeh Askari, Mobin Bagherian, Sadegh Mohammadian, Mohammad Izadi, Mahdieh Soleymani Baghshah

ICML 2026 · Vision-language and multimodal › Cross-modal learning · approach: Contrastive vision-language alignment

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add aligned symbol cues to both image and prompt (vs. text-only or image-only cues) (ablation)Qwen2.5-VL-7B · Synthetic scene description, 20 objectsF1: 0.21 -> 0.62Improved
add cues to the image only (ablation)Qwen2.5-VL-7B · Synthetic scene description, 10 objectsaccuracy: 0.42 -> 0.38Hurt
add structured cues with Grounding IDs instead of VISERQwen2.5-VL-7B · Countingaccuracy: 43.33 -> 53.00Improved
add structured cues to image and promptLLaVA-1.5 · MS-COCO CHAIRCHAIR_s: 51.60 -> 41.00Improved
add structured cues to image and promptGPT-4o · MS-COCO CHAIRCHAIR_s: 29.20 -> 23.20Improved
use row-level symbols with lines rather than per-object symbols (ablation)Qwen2.5-VL-7B · Synthetic variants, 20 objectsaccuracy: 0.19 -> 0.40Improved
use row separator lines in addition to symbols (ablation)Qwen2.5-VL-7B · Synthetic variants, 20 objectsaccuracy: 0.33 -> 0.40Improved
use 4x4 grid layout instead of row partitions on synthetic images (ablation)Qwen2.5-VL-7B · Synthetic variants, 20 objectsaccuracy: 0.40 -> 0.22Hurt
add structured cues on POPEQwen2.5-VL · POPE Randomaccuracy: 89.73 -> 85.67Hurt
add structured cues on POPELLaVA-1.5 · POPE Randomaccuracy: 87.27 -> 89.23Improved
ensemble five partition layouts with majority votingQwen2.5-VL structured · POPE Adversarialaccuracy: 83.73 -> 85.72Improved
swap object activations between contexts (causal test)Qwen2.5-VL-7B · Activation swapaccuracy: standard 1.00 -> 0.02; swap accuracy 0.98Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy