AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

GUI-AIMA: Aligning Intrinsic Multimodal Attention with a Context Anchor for GUI Grounding

Shijie Zhou, Viet Lai, Hao Tan, Jihyung Kil, Wanrong Zhu, Changyou Chen, Ruiyi Zhang

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Benchmarks and analysis of MLLMs · 7 citations (OpenAlex)

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 ANCHOR-token aggregation with uniform head weights (simplified attention grounding) over vanilla token-wise attention (ablation)GUI-AIMA (45k) · ScreenSpot-Proaccuracy: 37.51 -> 40.23Improved
weight heads with visual-sink top-1 query token (global) instead of uniform weights (ablation)GUI-AIMA (45k) · ScreenSpot-Proaccuracy: 40.23 -> 42.13Improved
weight heads with all query tokens (TAG style) instead of uniform (ablation)GUI-AIMA (45k) · ScreenSpot-Proaccuracy: 40.23 -> 37.63Hurt
weight heads with ANCHOR token only instead of uniform (ablation)GUI-AIMA (45k) · ScreenSpot-v2accuracy: 88.92 -> 87.81Hurt
use top-3 instead of top-1 global visual-sink tokens (ablation)GUI-AIMA (45k) · ScreenSpot-v2accuracy: 89.39 -> 84.98Hurt
use layer-wise instead of global shared visual-sink token (ablation)GUI-AIMA (45k) · ScreenSpot-v2 / ScreenSpot-Proaccuracy: SSv2 89.39 -> 89.86; SSPro 42.13 -> 41.43Mixed
add overlap- and distance-aware weighted patch labeling (ablation)GUI-AIMA (45k) · ScreenSpot-Proaccuracy: 42.13 -> 43.39Improved
add two-step zoom-in inferenceGUI-AIMA-3B · ScreenSpot-Proaccuracy: 53.8 -> 61.5Improved
add zoom-in inferenceGUI-AIMA-3B · OSWorld-G-Refineaccuracy: 62.8 -> 68.1Improved
crop and zoom second step by 2x versus crop-only (ablation)GUI-AIMA two-step inference · ScreenSpot-Pro offset errorsimproved predictions: 49 -> 182Improved
apply GUI-AIMA to InternVL3.5-4B backbone (ablation)InternVL3.5-4B · ScreenSpot-v2accuracy: 85.1 -> 88.8Improved
stop gradient on head weights (ablation)GUI-AIMA-3B-lite (254k) · ScreenSpot-Proaccuracy: 49.78 -> 45.73Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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