GuirlVG: Incentivize GUI Visual Grounding via Empirical Exploration on Reinforcement Learning

Graphical user interface visual grounding (GUI-VG)-a core capability for GUI agents-has primarily relied on supervised fine-tuning (SFT) of multimodal large language models (MLLMs), demanding extensive data curation and significant training costs. However, as MLLMs continue to advance and even cover GUI domains during pretraining, the necessity of exhaustive SFT post-training becomes increasingly questionable. Meanwhile, the recent successes of rule-based reinforcement fine-tuning (RFT) suggest a more efficient alternative. However, despite its promise, the optimal manner of RFT for GUI-VG remains unexplored. To bridge this gap, we introduce GuirlVG, a reinforcement learning-based GUI-VG method built on a systematic empirical study and a novel stabilization technique. Preliminarily, we find that naive application of RFT underperforms the SFT baseline, motivating a deeper exploration of RFT. First, we decompose RFT into its core components and analyze the optimal formulation of each. Second, as part of this exploration, we propose a novel Adversarial KL Factor that dynamically stabilizes training to mitigate reward over-optimization. Third, we further explore the training configurations of RFT to enhance the effectiveness. Extensive experiments show that GuirlVG, with only 5.2K training samples, outperforms SFT methods trained on over 10M samples, achieving a +7.7% improvement on ScreenSpot, a +17.2% improvement on ScreenSpotPro and 91.9% accuracy on ScreenSpotV2. RFT (Trivial) 79.2% 83.4% + Tune β + Adv. KL + Img. Res. Figure 1: Step-by-step exploration of GuirlVG. Starting from trivial RFT, we progressively add Soft Reward Function, In-Bbox reward with point prediction, β tuning, our Adversarial KL Factor, image resolution prompting, and extended training. With only 5.2K data, GuirlVG surpasses SFT methods trained on up to 13.58M data. Circle size reflects data scale used by each method. Preprint. Under review.