MobileExplorer: Accelerating On-Device Inference for Mobile GUI Agents via Online Exploration
Runxi Huang, Liyu Zhang, Shengzhong Liu, Xiaomin Ouyang · 2026
Read MobileExplorer: Accelerating On-Device Inference for Mobile GUI Agents via Online Exploration by Runxi Huang, Liyu Zhang, Shengzhong Liu, Xiaomin Ouyang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-26546.txt.
Description
Runxi Huang and three colleagues speed on-device inference for mobile GUI agents with MobileExplorer. Cloud models raise privacy and latency worries, yet on-device agents remain underexplored. Their framework exploits slow vision-language reasoning with lightweight parallel interface exploration, storing traces as structured memory. A two-level rollback restores the initial state when naive backtracking fails. Traces compress into concise hints injected into the next prompt. On AndroidWorld and harder custom tasks, reasoning steps and latency fall 23 percent, with success rates holding or rising 5 percent.
Catalog record
| CREATOR | Runxi Huang, Liyu Zhang, Shengzhong Liu, Xiaomin Ouyang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54514 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.26546