Anticipate and Learn: Unleashing Idle-Time Compute in Proactive Agents

Haoyi Hu, Qirong Lyu, Xianghan Kong, Weiwen Liu, Jianghao Lin, Zixuan Guo, Yan Xu, Yasheng Wang, Weinan Zhang, Yong Yu

arXiv:2605.25971 · 2026-05-26 공개 · arXiv · PDF

task-completion hallucination-reduction persistent-memory dialogue-history membench proact proacteval proactive-agents

Abstract

While AI agents demonstrate remarkable capabilities in reasoning and tool use, they remain fundamentally reactive: they compute responses only after explicit user prompts. This paradigm ignores a critical opportunity: the idle time between interactions is largely wasted, leaving agents unable to prepare for future user needs. To bridge this gap, we introduce ProAct, a proactive agent architecture that leverages idle-time compute to anticipate and fulfill likely upcoming user needs. By analyzing evolving dialogue history together with persistent memory, ProAct predicts upcoming needs and iteratively acquires information, allowing the agent to resolve knowledge gaps and prepare evidence before the user initiates a query.To rigorously evaluate proactive capabilities, we also introduce ProActEval, a comprehensive benchmark comprising 200 scenarios across 40 domains, featuring predictable need chains and diverse user cognitive profiles. Empirical results demonstrate significant advantages over reactive baselines. ProAct accelerates task completion by reducing required turns by 14.8%, decreases user effort by 11.7%, and cuts hallucination rates by 28.1% on ProActEval. Furthermore, MemBench evaluations confirm that ProAct achieves state-of-the-art reflective accuracy, underscoring its sustained and robust performance.

한국어 요약

📋 한 줄 요약

**[Proactive Agent / Idle-Time Compute]** ProAct가 대화 사이 idle time에 대화 이력·persistent memory로 사용자 needs를 예측해 사전 정보 수집, ProActEval(200 시나리오·40 도메인)에서 turn 14.8%↓·effort 11.7%↓·hallucination 28.1%↓.

🎯 핵심 기여도

💡 핵심 아이디어

사용자 대화 사이 idle time은 낭비된 자원이며, 대화 이력과 persistent memory를 분석해 다가올 need를 예측하고 background에서 정보를 미리 수집하면 reactive agent의 latency·effort·hallucination을 동시에 줄일 수 있다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: Reactive→proactive로 agent 패러다임 전환 제안, idle-time compute라는 미활용 자원의 가치 정량 입증, ProActEval로 평가 표준 정립, hallucination 28% 감소는 신뢰성 관점에서 큰 가치. **한계**: 200 시나리오·40 도메인 합성 평가의 실세계 격차, persistent memory 의존성 — 메모리 부정확 시 잘못된 anticipation 위험, idle-time compute의 cost 고려 필요.

🚀 실용적 활용