On the Effects of Data Scale on UI Control Agents
Wei Li, William Bishop, Alice Li, Chris Rawles, Folawiyo Campbell-Ajala, Divya Tyamagundlu, Oriana Riva
- 🏛 Institutions
- Google DeepMind, Google
- 📅 Date
- June 6, 2024
- 📑 Publisher
- NeurIPS 2024 Datasets and Benchmarks Track
- 💻 Env
- Mobile
- 🔑 Keywords
TLDR
Studies how UI-control agent performance scales with more fine-tuning data and releases AndroidControl, a dataset of over 15K demonstrations across 833 Android apps. The paper shows strong in-domain scaling trends while highlighting that out-of-domain generalization remains harder.
Related papers
- UI-R1: Enhancing Efficient Action Prediction of GUI Agents by Reinforcement LearningMarch 27, 2025 · arXiv
- Explorer: Scaling Exploration-driven Web Trajectory Synthesis for Multimodal Web AgentsJuly 2025 · Findings of ACL 2025
- Don't Act Blindly: Robust GUI Automation via Action-Effect Verification and Self-CorrectionApril 7, 2026 · ACL 2026
- PSPA-Bench: A Personalized Benchmark for Smartphone GUI AgentMarch 31, 2026 · arXiv
- Video-Based Reward Modeling for Computer-Use AgentsMarch 10, 2026 · arXiv
- SecAgent: Efficient Mobile GUI Agent with Semantic ContextMarch 9, 2026 · arXiv