Perceive Before Reasoning: A Pre-Reasoning Perception Framework for Efficient and Reliable Proactive Mobile Agents

2026-06-02

AI Summary

This work addresses the challenges of goal misalignment and redundant reasoning arising from joint decision-making on intervention timing and assistance modality in proactive mobile assistants. To mitigate these issues, the authors propose a two-stage decoupled framework: a lightweight multimodal proactive perceiver (MPP) first determines whether intervention is necessary and compresses contextual information, activating the proactive agent reasoner (PAR) only when needed to generate appropriate assistance. By decoupling intervention decision from assistance generation and introducing a perception-frontloading mechanism, the approach significantly reduces false triggers while simultaneously improving task success rates and reasoning efficiency, as demonstrated on the ProactiveMobile benchmark.

Abstract

Multimodal large language models (MLLMs) have substantially advanced mobile agents, yet proactive mobile assistance remains challenging because agents must decide \emph{when} to intervene before determining \emph{how} to assist. Existing systems often implement these two decisions within a unified MLLM-based pipeline, leading to goal misalignment between conservative intervention filtering and comprehensive assistance generation, as well as redundant inference when the agent should remain silent. To address these limitations, we propose the \textbf{Pre-Reasoning Perception Framework (PRPF)}, a two-stage framework built on perceiving before reasoning. PRPF introduces a lightweight Multimodal Proactive Perceptor (MPP) for intervention gating and context compression, and activates the Proactive Agent Reasoner (PAR) only when intervention is warranted. Experiments on the ProactiveMobile benchmark show that PRPF substantially reduces false trigger rates (FTR) while improving success rates (SR) and inference efficiency over the ProactiveMobile baseline.

Problem

• proactive mobile agents
• intervention timing
• multimodal large language models
• goal misalignment
• redundant inference

Innovation

• Pre-Reasoning Perception Framework
• Multimodal Proactive Perceptor
• Proactive Mobile Agents
• Intervention Gating
• Two-Stage Architecture