Variational-Auto-Encoders

• This repository contains a Variational Autoencoder (VAE) model for anomaly detection, specifically for credit card fraud detection.
• It was developed as part of my M.Sc. Computer Science dissertation.
• While this implementation is niche, it demonstrates how generative models like VAEs can be applied to anomaly detection.
• This project is intended solely for academic purposes, exploring the potential of VAEs in fraud detection and anomaly identification.

Tech Stack

• Python (Programming Langauge)
• TensorFlow (Framework)
• PyTorch (Framework)

Note Books

• TensorFlow Implementation: The original VAE model was implemented using TensorFlow.
• PyTorch Implementation: A new notebook has been added, featuring the same model reimplemented in PyTorch.

Dataset

The Dataset used for this project was solely from the Kaggle repository. The link below takes you straight to the repo: https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud.