arXiv cs.CL· Bo Qu, Zhurong Wang, Daisuke Yagi, Zhen Xu, Yang Zhao, Yinan Shan, Frank Zahradnik·· 11 小时前AI 评分31
LLM 增强的自演化强化学习用于多步电商支付欺诈风险检测
LLM-Enhanced Self-Evolving Reinforcement Learning for Multi-Step E-Commerce Payment Fraud Risk Detection
AI 导读
该研究将电商支付欺诈检测建模为多步马尔可夫决策过程，用 LLM 迭代优化强化学习的奖励函数，在真实数据上提升了欺诈检测准确率并展现零样本能力。长期评估验证了该 LLM 增强 RL 框架的有效性与鲁棒性，成果发表于 ACL 2025 工业赛道。
正文 · 原文
Abstract:This paper presents a novel approach to e-commerce payment fraud detection by integrating reinforcement learning (RL) with Large Language Models (LLMs). By framing transaction risk as a multi-step Markov Decision Process (MDP), RL optimizes risk detection across multiple payment stages. Crafting effective reward functions, essential for RL model success, typically requires significant human expertise due to the complexity and variability in design. LLMs, with their advanced reasoning and coding capabilities, are well-suited to refine these functions, offering improvements over traditional methods. Our approach leverages LLMs to iteratively enhance reward functions, achieving better fraud detection accuracy and demonstrating zero-shot capability. Experiments with real-world data confirm the effectiveness, robustness, and resilience of our LLM-enhanced RL framework through long-term evaluations, underscoring the potential of LLMs in advancing industrial RL applications.
| Comments: | 12 pages, 12 figures, ACL 2025 industry track |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2509.18719 [cs.LG] |
| (or arXiv:2509.18719v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2509.18719 arXiv-issued DOI via DataCite |
|
| Journal reference: | In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track), pages 92-103, 2025 |
| Related DOI: | https://doi.org/10.18653/v1/2025.acl-industry.9
DOI(s) linking to related resources |
Submission history
From: Bo Qu [view email]
[v1]
Tue, 23 Sep 2025 07:07:16 UTC (698 KB)
来源：arXiv cs.CL · arxiv.org