AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

FDC-Ground: Improving GRPO for GUI Grounding via Exponential Rewards and Fact-Aligned Pruning

Xiangjian Zeng, Wenjing Li, Qingqiang Wu, Liang Zhang

AAAI 2026 · Reinforcement learning › RL from feedback · approach: RL fine-tuning of generative models · 0 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
replace sparse binary reward with EDDR (ablation)GRPO · ScreenSpotaverage accuracy (%): 74.4 -> 85.9Improved
replace square decayed distance reward with EDDR (ablation)GRPO · ScreenSpotaverage accuracy (%): 85.5 -> 85.9Improved
add top-m selection by absolute advantage (ablation)EDDR + GRPO · ScreenSpotaverage accuracy (%): 85.9 -> 86.7Improved
add fact-aligned filtering to top-m pruning (ablation)EDDR + Top-m · ScreenSpotaverage accuracy (%): 86.7 -> 88.1Improved
apply FDC-Pruning (g=4, m=2) versus plain GRPO (ablation)GRPO · ScreenSpot Web subset (Qwen2.5-VL-3B)average accuracy (%): 66.1 -> 72.0Improved
apply FDC-Pruning (g=4, m=2) versus plain GRPO (ablation)GRPO · ScreenSpot Web subset (Qwen2.5-VL-3B)training time (s): 11541 -> 9789Improved
increase group size from 4 to 16 (ablation)FDC-Pruning · ScreenSpot Web subset (Qwen2.5-VL-3B)average accuracy (%): 72.0 -> 76.6Improved
increase group size from 16 to 64 (ablation)FDC-Pruning · ScreenSpot Web subset (Qwen2.5-VL-3B)average accuracy (%): 76.6 -> 74.8Hurt
increase group size from 16 to 64 (ablation)FDC-Pruning · ScreenSpot Web subset (Qwen2.5-VL-3B)training time (s): 19675 -> 57702Hurt
train with FDC-Ground RLQwen2.5-VL-3B · ScreenSpot-v2average accuracy (%): 46.9 -> 85.5Improved
add entropy regularizationFDC-Ground · GUI grounding trainingreward plateauNo clear effect
add exploration-oriented rewardsFDC-Ground · GUI grounding trainingaccuracyHurt

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