FairnessExplainabilityAccountabilitycs.LG
Reinforcement-Guided Hyper-Heuristic Hyperparameter Optimization for Fair and Explainable Spiking Neural Network-Based Financial Fraud Detection
Sadman Mohammad Nasif, Md Abrar Jahin, M. F. Mridha
August 23, 2025
Knowledge-Based Systems
2 citations
Key Takeaways
-
Fair AI can be technically achievable in high-stakes applications: The system maintained over 98% predictive equality across demographic groups while detecting 90.8% of fraud cases, demonstrating that fairness constraints don't necessarily require sacrificing detection performance in financial fraud scenarios.
-
Interpretability matters for fraud detection: The framework incorporates explainable AI through saliency maps and spike activity profiling, allowing stakeholders to understand why a transaction was flagged as fraudulent—a critical requirement for customer trust and regulatory compliance in banking systems.
-
Energy efficiency is a practical benefit of responsible AI design: The sparse architecture of the spiking neural network approach reduces energy consumption compared to traditional dense AI models, showing that building fair and explainable systems can also yield environmental and operational cost benefits.
-
Automated optimization can help balance competing priorities: The reinforcement learning-based optimizer (RHOSS) adaptively selects parameters while respecting both fairness and recall constraints simultaneously, suggesting that responsible AI doesn't require manual trade-offs between accuracy, fairness, and other objectives—automation can help navigate these tensions.
-
Computational costs of responsible AI are manageable: While the optimization process adds upfront computational expense, the authors note this cost is "amortized at deployment," indicating that the infrastructure burden of building fair and explainable systems may be acceptable when spread across operational use.
Abstract
The growing adoption of home banking systems has increased cyberfraud risks, requiring detection models that are accurate, fair, and explainable. While AI methods show promise, they face challenges including computational inefficiency, limited interpretability of spiking neural networks (SNNs), and instability in reinforcement learning (RL)-based hyperparameter optimization. We propose a framework combining a Cortical Spiking Network with Population Coding (CSNPC) and a Reinforcement-Guided Hyper-Heuristic Optimizer (RHOSS). CSNPC leverages population coding for robust classification, while RHOSS applies Q-learning to adaptively select low-level heuristics under fairness and recall constraints. Integrated within the MoSSTI framework, the system incorporates explainable AI via saliency maps and spike activity profiling. Evaluated on the Bank Account Fraud (BAF) dataset, the model achieves 90.8% recall at 5% false positive rate, outperforming prior spiking and classical models while maintaining over 98% predictive equality across demographic groups. Although RHOSS introduces offline optimization cost, it is amortized at deployment. The sparse architecture of CSNPC further reduces energy consumption compared to dense ANNs. Results demonstrate that combining population-coded SNNs with RL-guided hyper-heuristics enables fair, interpretable, and high-performance fraud detection.
Paper Info
arXiv ID2508.16915
PublishedAugust 23, 2025
RevisedMarch 23, 2026
Citations
2
Relevance82%
VenueKnowledge-Based Systems
TypeJournalArticle
Categorycs.LG
Responsible AI Topics
FairnessExplainabilityAccountability
Links
Citation data from Semantic Scholar