Research map / Multi-agent systems and games / LLM agents
UIPro: Unleashing Superior Interaction Capability For GUI Agents
Hongxin Li, Jingran Su, Jingfan Chen, Zheng Ju, Yuntao Chen, Qing Li, Zhaoxiang Zhang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add denoising of GUI understanding data (ablation) | UIPro-SLiME · FuncPred | grounding accuracy: 64.3 -> 66.7 | Improved |
| add denoising of GUI understanding data (ablation) | UIPro-SLiME · RefExp | grounding accuracy: 43.5 -> 45.8 | Improved |
| add denoising of GUI understanding data (ablation) | UIPro-SLiME · VWB AG | grounding accuracy: 32.0 -> 34.0 | Improved |
| add 20.6M GUI understanding pre-training data (vs none) (ablation) | UIPro-SLiME · OmniAct-Desktop | Step SR: 15.4 -> 25.1 | Improved |
| add 5.9M GUI understanding pre-training data (vs none) (ablation) | UIPro-SLiME · OmniAct-Desktop | Step SR: 15.4 -> 18.9 | Improved |
| scale pre-training data from 5.9M to 20.6M (ablation) | UIPro-SLiME · OmniAct-Desktop | Step SR: 18.9 -> 25.1 | Improved |
| add unified action space data sources to fine-tuning (ablation) | UIPro-7B · AndroidControl-High (Swipe) | action accuracy: 22.6 -> 50.9 | Improved |
| add unified action space data sources to fine-tuning (ablation) | UIPro-7B · AndroidControl-High (Wait) | action accuracy: 43.6 -> 73.4 | Improved |
| add unified action space data sources to fine-tuning (ablation) | UIPro-7B · GUIAct (Answer) | action accuracy: 12.6 -> 43.0 | Improved |
| unify action spaces when mixing data sources (ablation) | UIPro · AITW/AndroidControl/GUIAct | Step SR | Improved |
| add pre-training on GUI understanding data (ablation) | UIPro · GUI agent benchmarks | Step SR | Improved |
| use UIPro-Qwen2VL instead of OS-ATLAS | UIPro-Qwen2VL-7B · AITW | Overall step SR: 63.1 -> 70.4 | Improved |
Closest papers by meaning
- ShowUI: One Vision-Language-Action Model for GUI Visual Agent (CVPR 2025 · similarity 0.91)
- AssistGUI: Task-Oriented PC Graphical User Interface Automation (CVPR 2024 · similarity 0.90)
- Aguvis: Unified Pure Vision Agents for Autonomous GUI Interaction (ICML 2025 · similarity 0.90)
- UI-Vision: A Desktop-centric GUI Benchmark for Visual Perception and Interaction (ICML 2025 · similarity 0.90)
- MMBench-GUI: A Unified Hierarchical Evaluation Framework for Multi-Platform GUI Agents (CVPR 2026 · similarity 0.89)
- GUI-Xplore: Empowering Generalizable GUI Agents with One Exploration (CVPR 2025 · similarity 0.89)
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.
