AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Reason with Thumbnails, Answer with Focus: An Efficient and Effective Paradigm for Multimodal Grounded Visual Reasoning

An-Lan Wang, Guozhi Tang, Lei Liao, Hanshen Zhu, Kai Huang, Jingqun Tang, Jiaming Zhou, Kun-Yu Lin

ICML 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Attention-based visual token pruning

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes an efficient grounded visual reasoning paradigm for multimodal LLMs that reasons over thumbnails and answers with focused regions, aiming to cut the cost of multi-round ground-then-answer methods (abstract only; details not available).

Problem

Grounded Visual Reasoning (GVR) improves interpretability of multimodal LLMs by selecting relevant image regions before answering, but its multi-round ground-then-answer and reasoning process is much more computationally expensive than non-GVR methods.

Approach

Reason over low-resolution thumbnails and answer with focused high-resolution regions to get efficient and effective GVR; the abstract does not give further details of the mechanism.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add RTAF paradigm (thumbnail reasoning, high-res crop answering) to Ground-R1 (ablation)Ground-R1 · VisCoT test averageaverage visual tokens: 847 -> 627Improved
add RTAF paradigm to Ground-R1 (ablation)Ground-R1 · VisCoT test averageaccuracy: 77.8 -> 76.2Hurt
add new reward functions to RTAF (ablation)RTAF · VisCoT test averageaccuracy: 76.2 -> 77.1Improved
add new reward functions to RTAF (ablation)RTAF · VisCoT test averageaverage visual tokens: 627 -> 553Improved
add training data selection strategy (ablation)RTAF with rewards · VisCoT test averageaccuracy: 77.1 -> 78.3Improved
add bounding box reward (ablation)RTAF · VisCoT test averageaccuracy: 76.2 -> 76.7Improved
add calibrated efficiency reward (ablation)RTAF · VisCoT test averageaccuracy: 76.7 -> 77.1Improved
double the efficiency reward weight (2 x Reff) (ablation)RTAF · VisCoT test averageaccuracy: 77.1 -> 74.5Hurt
use class-wise data selection (Combination 3) instead of random data (ablation)RTAF · VisCoT test averageaccuracy: 77.1 -> 78.3Improved
use RTAF instead of Qwen2.5-VL-7BQwen2.5-VL-7B · VisCoT test averageaccuracy: +5.8% accuracy, 471 vs. 391 tokensImproved
use RTAF instead of Ground-R1 on HR-Bench-4K token usagePixel Reasoner and DeepEyes · HR-Bench-4Kaverage visual tokens: 13,118 vs 41,080 and 39,418Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 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