CORE: Reducing UI Exposure in Mobile Agents via Collaboration Between Cloud and Local LLMs

Gucongcong Fan, Chaoyue Niu, Chengfei Lyu, Fan Wu, Guihai Chen

🏛 Institutions: SJTU, Alibaba Group
📅 Date: October 17, 2025
📑 Publisher: NeurIPS 2025 (Poster)
💻 Env: Mobile
🔑 Keywords: framework privacy cloud-local collaboration UI exposure reduction CORE

TLDR

CORE studies how to reduce unnecessary screen exposure when mobile agents depend on cloud LLMs for planning and action selection. It partitions the UI into layout-aware blocks and lets local and cloud models collaborate on planning and decision-making so only task-relevant UI subsets are sent to the cloud, substantially reducing exposure while keeping accuracy close to cloud-only systems.

Open paper arXiv Edit on GitHub Report issue