@inproceedings{liu-etal-2026-learnact,
    title = "{L}earn{A}ct: Few-Shot Mobile {GUI} Agent with a Unified Demonstration Benchmark",
    author = "Liu, Guangyi  and
      Zhao, Pengxiang  and
      Liu, Liang  and
      Chen, Zhiming  and
      Chai, Yuxiang  and
      Liang, Yaozhen  and
      Wang, WenHao  and
      Chen, Siheng  and
      Lu, Zhengxi  and
      Ren, Shuai  and
      Wang, Hao  and
      He, Shibo  and
      Liu, Yong  and
      Meng, Wenchao",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.1491/",
    pages = "29820--29843",
    ISBN = "979-8-89176-395-1",
    abstract = "Mobile GUI agents show promise in automating tasks but face significant generalization challenges in long-tail scenarios. While learning from few-shot demonstrations is an emerging solution, its progress is hindered by two critical gaps: the lack of a comprehensive benchmark for systematic evaluation on mobile devices, and the absence of a systematic framework designed to learn from demonstrations in this domain. To address these gaps, we introduce \textbf{LearnGUI}, the first comprehensive benchmark designed for studying demonstration-based learning in mobile agents, comprising 2,252 offline and 101 online tasks. We further develop \textbf{LearnAct}, a modular agent framework engineered to systematically extract, retrieve, and leverage knowledge from visual demonstrations. Extensive evaluations across six backbone models validate our approach: LearnAct achieves dramatic improvements for general-purpose models (e.g., Gemini-2.5-Pro: 38.5{\%}{\textrightarrow}58.9{\%}) and specialized models alike (e.g., UI-TARS-7B-SFT{'}s online success rate: 18.1{\%}{\textrightarrow}32.8{\%}), demonstrating consistent gains across model architectures. Our work provides a robust benchmark and a systematic framework, paving the way for more adaptable and practical mobile agents. Our code and data are publicly available at \url{https://lgy0404.github.io/LearnAct/}."
}