---
id: 20261103-T0-07
title: "不复制环境也能扩展Agent测试时算力，新方法用世界模型做预测"
title_en: "Agent Test-Time Scaling Without Environment Replicas via World Modeling"
url: https://ai.daily.yangsir.net/daily/20261103-T0-07
issue_date: 2026-11-03
publish_date: 2026-11-02T05:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2610.40927
---

# 不复制环境也能扩展Agent测试时算力，新方法用世界模型做预测

arXiv 论文提出一种面向 Agent 测试时扩展（TTS）的世界模型方法。常见 TTS 做法是并行采样多条候选轨迹再筛选，但现实中轨迹往往无法并行执行，这一假设在部署场景下不成立。该研究用 agentic world modeling 绕开环境复制，让 Agent 在无法并行跑轨迹时仍能做测试时扩展。具体效果数据摘要中未给出。

## English Version

**Agent Test-Time Scaling Without Environment Replicas via World Modeling**

An arXiv paper proposes an agentic world modeling approach for agent test-time scaling (TTS). Common TTS methods sample and select among multiple candidate trajectories in parallel, but that assumption breaks in realistic deployments where trajectories cannot be executed concurrently. The work avoids environment replication through world modeling, enabling test-time scaling when parallel rollouts are impossible. Specific performance results were not provided in the abstract.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2610.40927)

**详情页**：https://ai.daily.yangsir.net/daily/20261103-T0-07

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*