mlsanityData to 14 Oct 2026

Behavior-Grounded Semantic Enrichment for Financial Fraud Modeling and Reasoning

arXiv 2609.34211 ↗First posted 28 Sep 2026cs.AIcited by 0

On the map: LLM reasoning → Reasoning for agents, planning and code

Abstract (arXiv)

In financial fraud detection, rich semantic context can provide important evidence for transaction behavior modeling and fraud reasoning. However, public real-world financial datasets often lack rich semantics due to privacy constraints. Consequently, synthetic datasets incorporate generated semantics, but at the cost of behavioral realism; textual descriptions for contextual reasoning remain scarce. We address this gap through a semantic enrichment framework grounded in original transaction behavior to simulate multimodal financial data. We (1) propose a multi-agent semantic enrichment framework that generates interpretable financial semantics grounded in transaction behavior through role-specialized agents and consistency refinement, and (2) newly contribute a valuable multimodal financial fraud dataset, MS-FFSD, enriched with structured semantics and textual semantics while preserving real-data-grounded transaction behavior. Furthermore, we systematically analyze the quality and utility of semantic enrichment. Results demonstrate statistical fidelity and framework generalizability, while showing that richer semantics benefit fraud modeling and context-aware LLM reasoning. Overall, this work advances multimodal financial fraud research and bridges emerging LLM and multi-agent capabilities with operational anti-fraud practice. The framework and dataset are released at https://github.com/AI4Risk/MS-FFSD.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning. Builds on 32 of its 58 references that are in the map. First posted on arXiv 28 Sep 2026.

Builds on

32 of 58 references.

References · 32

  1. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  2. 2017Graph Attention Networkscited by 8,923
  3. 2025Qwen3 Technical Reportcited by 8,820
  4. 2017Inductive Representation Learning on Large Graphscited by 6,320
  5. 2022Large Language Models are Zero-Shot Reasonerscited by 4,374
  6. 2025Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Modelscited by 1,261
  7. 2023A Survey on Evaluation of Large Language Modelscited by 1,242
  8. 2024MM-LLMs: Recent Advances in MultiModal Large Language Modelscited by 278
  9. 2020Enhancing Graph Neural Network-based Fraud Detectors against Camouflaged Fraudsterscited by 250
  10. 2021A Comprehensive Survey on Graph Anomaly Detection with Deep Learningcited by 230
  11. 2022Rethinking Graph Neural Networks for Anomaly Detectioncited by 187
  12. 2024The Revolution of Multimodal Large Language Models: A Surveycited by 137
20 more
  1. 2020Logical Natural Language Generation from Open-Domain Tablescited by 90
  2. 2020Tabular Transformers for Modeling Multivariate Time Seriescited by 77
  3. 2022DGraph: A Large-Scale Financial Dataset for Graph Anomaly Detectioncited by 63
  4. 2023Realistic Synthetic Financial Transactions for Anti-Money Laundering Modelscited by 50
  5. 2023TabuLa: Harnessing Language Models for Tabular Data Synthesiscited by 42
  6. 2024Semi-supervised Credit Card Fraud Detection via Attribute-Driven Graph Representationcited by 26
  7. 2024DataNarrative: Automated Data-Driven Storytelling with Visualizations and Textscited by 25
  8. 2024Graph Neural Networks for Financial Fraud Detection: A Reviewcited by 25
  9. 2024Partitioning Message Passing for Graph Fraud Detectioncited by 23
  10. 2023Label Information Enhanced Fraud Detection against Low Homophily in Graphscited by 22
  11. 2023Demystifying Fraudulent Transactions and Illicit Nodes in the Bitcoin Network for Financial Forensicscited by 19
  12. 2025SpaceGNN: Multi-Space Graph Neural Network for Node Anomaly Detection with Extremely Limited Labelscited by 17
  13. 2022PLOG: Table-to-Logic Pretraining for Logical Table-to-Text Generationcited by 16
  14. 2019Synthesizing Credit Card Transactionscited by 10
  15. 2024ControlMath: Controllable Data Generation Promotes Math Generalist Modelscited by 9
  16. 2024Efficient Dynamic Attributed Graph Generationcited by 5
  17. 2025PANTHER: Generative Pretraining Beyond Language for Sequential User Behavior Modelingcited by 4
  18. 2025ReTabAD: A Benchmark for Restoring Semantic Context in Tabular Anomaly Detectioncited by 3
  19. 2025Enhancing Foundation Models in Transaction Understanding with LLM-based Sentence Embeddingscited by 2
  20. 2026Learning to Generate and Extract: A Multi-Agent Collaboration Framework For Zero-shot Document-level Event Arguments Extractioncited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Semantic Pareto-DQN: A Multi-Objective Reinforcement Learning Framework for Financial Anomaly Detection
  2. 2026SAGE: An LLM-driven Self Reflective Agentic Framework for Fraud Detection
  3. 2026Reinforcement Learning of Large Language Models for Interpretable Credit Card Fraud Detection
  4. 2025Exploring the In-Context Learning Capabilities of LLMs for Money Laundering Detection in Financial Graphs
  5. 2025Understanding Structured Financial Data with LLMs: A Case Study on Fraud Detection
  6. 2025When AI Agents Collude Online: Financial Fraud Risks by Collaborative LLM Agents on Social Platforms
  7. 2023Beyond Classification: Financial Reasoning in State-of-the-Art Language Models
  8. 2025FAA Framework: A Large Language Model-Based Approach for Credit Card Fraud Investigations