AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

InfiGUI-G1: Advancing GUI Grounding with Adaptive Exploration Policy Optimization

Yuhang Liu, Zeyu Liu, Shuanghe Zhu, Pengxiang Li, Congkai Xie, Jiasheng Wang, Xueyu Hu, Xiaotian Han, Jianbo Yuan, Xinyao Wang, Shengyu Zhang, Hongxia Yang and 1 more

AAAI 2026 · Reinforcement learning › RL from feedback · approach: GRPO stability and variants · 28 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 multi-answer generation to RLVR (ablation)InfiGUI-G1-7B · ScreenSpot-ProTop-1 accuracy: 47.6 -> 51.9Improved
use Adaptive Exploration Reward instead of naive reward (ablation)InfiGUI-G1-7B · ScreenSpot-ProTop-1 accuracy: 41.4 -> 51.9Improved
add rank factor k to AER (ablation)InfiGUI-G1-7B · ScreenSpot-ProTop-1 accuracy: 44.0 -> 51.9Improved
add collinear penalty (ablation)InfiGUI-G1-7B · ScreenSpot-ProTop-1 accuracy: 37.0 -> 51.9Improved
add collinear penalty (ablation)InfiGUI-G1-3B · ScreenSpot-ProTop-1 accuracy: 35.3 -> 45.2Improved
add AER to 3B model (ablation)InfiGUI-G1-3B · ScreenSpot-ProTop-1 accuracy: 38.4 -> 45.2Improved
add AEPO (multi-answer + AER) vs naive RLVRQwen2.5-VL-7B · ScreenSpot-ProTop-1 accuracy: 47.6 -> 51.9Improved
add AEPO vs naive RLVRQwen2.5-VL-3B · UI-VisionOverall accuracy: 19.4 -> 22.0Improved
add AEPO vs naive RLVRQwen2.5-VL-7B · ScreenSpot-Pro hard subsetaccuracy: 10.8 -> 17.4Improved
add AEPO vs naive RLVRQwen2.5-VL-3B · ScreenSpot-Pro hard subsetaccuracy: 25.5 -> 31.4Improved
single-pass multi-answer vs naive RLVR given multiple attemptsInfiGUI-G1-7B · ScreenSpot-Prosuccess rate: 52.1 (pass@4) -> 58.0Improved
AEPO on easy ScreenSpot-Pro samplesInfiGUI-G1 3B and 7B · ScreenSpot-Pro easy subsetaccuracy: 100 -> 100No clear effect

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