Explainable AI for FinTech (ExLiFT)
Transparent Machine Learning for Financial Services
FNR-funded doctoral research developing explainable AI techniques for clustering, classification, and resampling in machine learning pipelines for financial services.
- Programme
- FNR funding
- Format
- Industrial PhD
- Institution
- University of Luxembourg / SnT
Problem
Machine learning pipelines in financial services rely on clustering, classification, and resampling techniques for customer segmentation, risk scoring, fraud detection, and imbalanced-data handling. While these methods can achieve strong predictive performance, their outputs are often opaque to the risk managers, compliance teams, and regulators who depend on them.
In regulated environments, this lack of interpretability creates a practical barrier to adoption: models may perform well, but without understandable reasoning, they remain difficult to validate, audit, and justify in operational decision-making.
Approach
The ExLiFT project investigated how explainability techniques can be systematically applied across clustering, classification, and resampling workflows, making financial-services ML pipelines more transparent and auditable. The research was conducted at the University of Luxembourg / SnT under FNR funding, combining academic rigour with industry relevance.
Focus Areas
The work focused on interpretable customer segmentation, explainable risk and fraud modelling, and better visibility into class-imbalance handling, with the goal of supporting governance, model validation, and compliance-oriented review processes.
Relevance to Cognifinity
This research underpins Cognifinity's expertise in explainable AI for regulated financial environments, especially where transparency, auditability, and human review are as important as predictive performance.
Interested in XAI for Financial Services?
Get in touch to discuss how explainable AI can bring transparency and auditability to FinTech ML pipelines.
Contact R&D Team