AI pipeline enhances financial fraud detection with LLM-generated narratives

Researchers have developed an end-to-end pipeline for identifying money mule accounts, a critical component in combating financial fraud. The system utilizes a LightGBM classifier with 280 features, a TreeSHAP layer for explaining predictions, and a large language model (LLM) to generate natural-language narratives for analysts. This LLM-augmented approach significantly improved detection rates in a live deployment, increasing the yield from 61% to 89% and expanding true-positive coverage. AI

IMPACT This research demonstrates how LLMs can be integrated into financial systems to improve fraud detection and reduce analyst workload.

RANK_REASON The item is a research paper detailing a new method for detecting financial fraud using AI. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• Connected Papers
• DagsHub
• Gotit.pub
• Hugging Face
• LightGBM
• Litmaps
• LLM
• ScienceCast
• Scite
• TreeSHAP

• product
• paper

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →