Skip to content

ai-based-detection-hedge-fund-fraud

AI-Based Detection of Hedge Fund Fraud: A Systematic Survey and Research Agenda

View on GitHub

Information

Property Value
Language TeX
Stars 0
Forks 0
Watchers 0
Open Issues 0
License MIT License
Created 2026-02-13
Last Updated 2026-03-25
Last Push 2026-02-14
Contributors 1
Default Branch master
Visibility private

Datasets

This repository includes 12 dataset(s):

Dataset Format Size

| mlc_config.json | .json | 0.17 KB |

| questions.json | .json | 2.77 KB |

| questions.json | .json | 3.45 KB |

| questions.json | .json | 3.12 KB |

| questions.json | .json | 2.65 KB |

| questions.json | .json | 2.82 KB |

| questions.json | .json | 3.97 KB |

| questions.json | .json | 2.68 KB |

| questions.json | .json | 2.46 KB |

| questions.json | .json | 2.0 KB |

| questions.json | .json | 2.72 KB |

| charts.json | .json | 9.57 KB |

Reproducibility

This repository includes reproducibility tools:

  • Python requirements.txt

Status

  • Issues: Enabled
  • Wiki: Enabled
  • Pages: Enabled

README

AI-Based Detection of Hedge Fund Fraud

A Systematic Survey and Research Agenda

Compile Slides License: MIT

Overview

This repository showcases the paper "AI-Based Detection of Hedge Fund Fraud: A Systematic Survey and Research Agenda" by Joerg Osterrieder (Zurich University of Applied Sciences). The paper addresses a critical gap in the literature by providing the first systematic, hedge-fund-specific framework for AI-based fraud detection.

Key Numbers

Metric Value
Hedge fund AUM $4.5 trillion
Mean AUC degradation under adversarial attack 10.6%
Open research problems identified 10
Documented fraud cases available 50--100
Systematic literature corpus 105 papers

Three Contributions

  1. C1: Detection Pipeline Taxonomy -- A unified five-stage framework (data ingestion, feature engineering, model selection, explainability, deployment) mapping fraud types to AI methods
  2. C2: Adversarial and Regulatory Readiness -- Assessment of method robustness under adversarial attack and compliance with EU AI Act / SEC requirements
  3. C3: Research Roadmap -- Ten concrete open problems spanning data, methodology, and deployment challenges

Repository Structure

Section Topic Slides Quiz Charts
01 Introduction Slides Quiz 2
02 Background Slides Quiz 3
03 Detection Pipeline (C1) Slides Quiz 3
04 Literature Review Slides Quiz 2
05 Adversarial & Regulatory (C2) Slides Quiz 3
06 Research Agenda (C3) Slides Quiz 2
07 Conclusion Slides Quiz 2
08 Reproducibility Slides Quiz 2
A0 Search Protocol Slides Quiz 1
A1 Feature Engineering Slides Quiz 1
A2 Glossary Slides -- --

Technical Details

  • Format: Beamer slides (Madrid theme, 8pt, 16:9)
  • Charts: Standalone Python scripts generating PDF outputs
  • Paper: Full LaTeX source in paper/ directory

Folder Structure

ai-based-detection-hedge-fund-fraud/
├── 01_introduction/          # Section slides + quiz + charts
├── 02_background/
├── 03_detection_pipeline/
├── ...
├── A0_search_protocol/       # Appendix slides
├── A1_feature_engineering/
├── A2_glossary/
├── paper/                    # Full paper LaTeX source
├── template_beamer_final.tex # Shared Beamer template
├── notation.tex              # Shared notation macros
└── README.md

Requirements

  • Python 3.8+ with matplotlib, numpy, seaborn
  • LaTeX distribution with Beamer support

Installation

git clone https://github.com/Digital-AI-Finance/ai-based-detection-hedge-fund-fraud.git
cd ai-based-detection-hedge-fund-fraud
pip install -r requirements.txt

Citation

@article{osterrieder2025hedge,
  title={AI-Based Detection of Hedge Fund Fraud: A Systematic Survey and Research Agenda},
  author={Osterrieder, Joerg},
  year={2025},
  institution={Zurich University of Applied Sciences}
}

License

MIT License

(c) Joerg Osterrieder 2025