← Back to all news
arXiv Machine Learning By Zhengyu Fang, Liang Wang, Yingtong Dou, Yuzhong Chen, Shubham Jain, Junpeng Wang, Xiran Fan, Menghai Pan, Jiarui Sun, Yan Zheng, Yiwei Cai

PersonaAgent: Retrieval-Augmented Transaction Anomaly Scoring with Refreshable Multi-Entity Persona Memory

Read the original on arXiv Machine Learning →

PersonaAgent is a retrieval‑augmented transaction anomaly detection framework that stores the latest entity behavior signals as refreshable personas in memory. It updates these personas weekly via a batch job and, during inference, retrieves them with a lightweight adapter to augment a base anomaly classifier’s scoring. Benchmarks at Visa show that PersonaAgent can process 1,000 transactions in 1–2 ms while improving PR‑AUC by 0.65–6.58 % and production business metrics by 8.93–14.24 % relative to baseline models.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv AI
Sep 30

MAADBench: The Refreshable Paradigm for Anomaly Detection in Multi-Agent Systems

MAADBench is a refreshable benchmark for anomaly detection in multi‑agent systems powered by large language models. It addresses the challenge of keeping benchmarks current by sampling and coupling generative tasks, generating trace data under configurable LLM backbones, and automatically providing deterministic step‑level labels. The authors evaluated 25 anomaly‑detection methods on 5,200 labeled traces, finding that existing approaches depend heavily on supervision, struggle with subtle MAS‑specific anomalies, and lack robustness across different LLM backbones.

By Lei Ma, Dennis Hofmann, Haowen Xu, Joshua DeOliveira, Peter VanNostrand, Lei Cao, Elke Rundensteiner
llmsagentsbenchmarks
More like this →
arXiv Machine Learning
Sep 24

SR-Fraud: An Outcome-Supervised Reflective LLM Agent Framework for Non-Stationary Payment Fraud Detection

SR‑Fraud is a framework that uses a frozen, stateless LLM agent to score transactions in real time while an offline reflection agent proposes boundary hypotheses based on matured errors. The system then verifies these hypotheses deterministically before updating its knowledge state. On a production payment‑fraud benchmark, SR‑Fraud outperforms both static and periodically retrained CatBoost models and successfully detects an emerging fraud burst.

By Xuwei Tan, Yao Ma, Xueru Zhang
llmsagentsbenchmarks
More like this →
arXiv AI
Jun 17

An AI Security Agent for Banking: Multi-Vector Fraud and AML Detection Across Retail and Corporate Accounts

arXiv:2606. 17555v1 Announce Type: cross Abstract: Banks simultaneously face signature-based fraud (card-not-present attacks, account takeover, ATM cloning) and behavioural financial crime (structuring, layering, mule networks, business email compromise) -- two threat families with fundamentally different detection requirements.

By Joseph Walusimbi, Joshua Benjamin Ssentongo
agents
More like this →