arXiv cs.CL· Xingshan Zeng, Lingzhi Wang, Weiwen Liu, Liangyou Li, Yasheng Wang, Lifeng Shang, Xin Jiang, Qun Liu·· 6 天前AI 评分24
ARTIS：通过迭代仿真实现智能体风险感知的测试时扩展
ARTIS: Agentic Risk-Aware Test-Time Scaling via Iterative Simulation
AI 导读
针对智能体动作直接作用于外部环境、后果不可逆且代价高昂的问题，研究者提出 ARTIS 框架，通过在执行前用仿真交互进行测试时探索，将探索与承诺解耦，在不引入环境风险的前提下扩展推理时计算以提升动作级可靠性。
正文 · 原文
Abstract:Current test-time scaling (TTS) techniques enhance large language model (LLM) performance by allocating additional computation at inference time, yet they remain insufficient for agentic settings, where actions directly interact with external environments and their effects can be irreversible and costly. We propose ARTIS, Agentic Risk-Aware Test-Time Scaling via Iterative Simulation, a framework that decouples exploration from commitment by enabling test-time exploration through simulated interactions prior to real-world execution. This design allows extending inference-time computation to improve action-level reliability and robustness without incurring environmental risk. We further show that naive LLM-based simulators struggle to capture rare but high-impact failure modes, substantially limiting their effectiveness for agentic decision making. To address this limitation, we introduce a risk-aware tool simulator that emphasizes fidelity on failure-inducing actions via targeted data generation and rebalanced training. Experiments on multi-turn and multi-step agentic benchmarks demonstrate that iterative simulation substantially improves agent reliability, and that risk-aware simulation is essential for consistently realizing these gains across models and tasks.
| Comments: | Accepted by NeurIPS 2026 |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2602.01709 [cs.CL] |
| (or arXiv:2602.01709v3 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2602.01709 arXiv-issued DOI via DataCite |
Submission history
From: Xingshan Zeng [view email]
[v1]
Mon, 2 Feb 2026 06:33:22 UTC (476 KB)
[v2]
Tue, 3 Feb 2026 03:19:49 UTC (476 KB)
[v3]
Fri, 9 Oct 2026 01:49:03 UTC (457 KB)
来源：arXiv cs.CL · arxiv.org