Fraud is Not Just Rarity: A Causal Prototype Attention Approach to Realistic Synthetic Oversampling

2025-07-19

AI Summary

Credit card fraud detection suffers from severe class imbalance and latent, evolving fraud patterns. Existing oversampling techniques (e.g., SMOTE) and generative models often yield overconfident classifiers with insufficient inter-class separation. To address this, we propose the Causal Prototype Attention Classifier (CPAC), which integrates an interpretable prototype-based attention mechanism into a VAE-GAN framework. CPAC enables classifier-guided optimization of the latent space, thereby improving both synthetic sample fidelity and inter-class discriminability. By incorporating causal modeling, CPAC enforces semantic consistency of fraud prototypes, enhancing cluster coherence and generalization. Evaluated on standard benchmarks, CPAC achieves 93.14% F1-score and 90.18% recall—outperforming state-of-the-art baselines significantly. This work establishes a novel paradigm for imbalanced learning that jointly advances interpretability, robustness, and classification efficacy.

Technology Category

• Machine Learning: Causal Learning
• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Reasoning under Uncertainty: Causality

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Detecting fraudulent credit card transactions remains a significant challenge, due to the extreme class imbalance in real-world data and the often subtle patterns that separate fraud from legitimate activity. Existing research commonly attempts to address this by generating synthetic samples for the minority class using approaches such as GANs, VAEs, or hybrid generative models. However, these techniques, particularly when applied only to minority-class data, tend to result in overconfident classifiers and poor latent cluster separation, ultimately limiting real-world detection performance. In this study, we propose the Causal Prototype Attention Classifier (CPAC), an interpretable architecture that promotes class-aware clustering and improved latent space structure through prototype-based attention mechanisms and we will couple it with the encoder in a VAE-GAN allowing it to offer a better cluster separation moving beyond post-hoc sample augmentation. We compared CPAC-augmented models to traditional oversamplers, such as SMOTE, as well as to state-of-the-art generative models, both with and without CPAC-based latent classifiers. Our results show that classifier-guided latent shaping with CPAC delivers superior performance, achieving an F1-score of 93.14% percent and recall of 90.18%, along with improved latent cluster separation. Further ablation studies and visualizations provide deeper insight into the benefits and limitations of classifier-driven representation learning for fraud detection. The codebase for this work will be available at final submission.

Problem

• Addresses class imbalance in fraud detection data
• Improves latent space structure for better clustering
• Enhances fraud detection performance using prototype attention

Innovation

• CPAC uses prototype-based attention mechanisms
• Combines VAE-GAN encoder for better clustering
• Classifier-guided latent shaping improves performance