PIRA-Bench: A Transition from Reactive GUI Agents to GUI-based Proactive Intent Recommendation Agents

Yuxiang Chai, Shunye Tang, Han Xiao, Rui Liu, Hongsheng Li

🏛 Institutions: CUHK MMLab, Nankai University, Huawei Research
📅 Date: March 9, 2026
📑 Publisher: arXiv
💻 Env: Desktop Mobile
🔑 Keywords: benchmark proactive recommendation intent inference continuous screenshots PIRF multithreaded trajectories

TLDR

PIRA-Bench studies proactive GUI assistance, where an agent infers user intent from continuous visual streams instead of waiting for explicit commands. It focuses on long, noisy, interleaved trajectories with user-profile context, and introduces PIRF as a memory-aware baseline for proactive intent recommendation.

Open paper arXiv Edit on GitHub Report issue