SR-Fraud is presented, an outcome-supervised reflective LLM framework that decouples request-time decisions from offline adaptation, and improves all detection metrics over its frozen decision agent, obtains higher point estimates than static and periodically retrained CatBoost, and detects an emerging fraud burst.
Abstract
Real-time payment fraud detection is a non-stationary streaming prediction problem: adversaries adapt before supervised labels mature, and localized burst attacks can cause losses before retraining. Production systems typically rely on tabular classifiers and rules, which can struggle to capture these emerging sequential patterns before periodic retraining occurs. We present SR-Fraud, an outcome-supervised reflective LLM framework that decouples request-time decisions from offline adaptation. A frozen, stateless agent scores each transaction from a Hybrid Episodic Window to track behavioral shifts, while an offline reflection agent proposes boundary hypotheses from matured errors. A deterministic verifier then admits only supported hypotheses into an executable knowledge state. On a production payment-fraud benchmark, SR-Fraud improves all detection metrics over its frozen decision agent, obtains higher point estimates than static and periodically retrained CatBoost, and detects an emerging fraud burst.
Financial statement fraud detectors are estimated on historical enforcement data and applied to filings that arrive years later, yet they are almost always evaluated as though the data-generating process were stationary. This paper treats temporal distribution shift as a first-class design constraint. Using 141,463 fir...
Noah Campbell, Claire Wilson, Emma Thompson· International journal of inf...· 0 citations
Financial transaction monitoring is increasingly expected to identify anomalous and potentially fraudulent activity while transactions are still operationally actionable. This requirement is more demanding than conventional offline classification because the detector is embedded in an unbounded, distributed stream in w...
John Kwesi Erbynn, Doreen Appiah, A. Ampomah-Britwum· Asian Journal of Economics B...· 0 citations
Online transaction fraud is rare, adversarial and temporally structured, which makes it difficult for static classifiers and for deep models that consume every available attribute. This paper proposes a cognitive-driven machine learning framework that mimics how a human fraud analyst works: first screen the evidence th...
Production fraud-detection models learn from a censored slice of their own traffic: transactions a legacy system blocked never receive fraud labels, so models train on a biased, approved-only distribution. We extend LLM-guided reward design, the framework of Qu et al. (2025) in which an LLM evolves the reward of a two-...
Bo Qu, Zhu-Rong Wang, Frank Zahradnik et al.· 0 citations
Payment fraud causes financial losses and undermines users'trust in web payment services. Although fraud detection is a critical defense for payment security, complex and diverse fraudulent behaviors can still evade detection. Generating fraud samples for data augmentation can broaden detectors'coverage of different fr...
Hui-Lin He, W. Han, Lin-Bo Shao et al.· 0 citations
Reinforcement learning (RL) has recently emerged as a viable adaptive replacement for traditional rule-based and supervised methods for detecting fraud perpetrated against banks that operate online. Nevertheless, RL-based fraud detection systems have always been evaluated in static, non-aggressive environments; hence,...
N. Shreyaa, S. Asifa, Sibi Amaran et al.· 2026 10th International Conf...· 0 citations
Research Manager Mike Walker reflects on a career spanning hacker culture, cybersecurity, and AI, sharing lessons about trust, teamwork, leadership, and why solving difficult problems requires curiosity, judgment, agility—and an understanding of the rules. The post Learn the rules: The hacker mindset appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduNov 17, 2026
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.
Microsoft Research Blog· microsoft.comNov 16, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.