Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Empowering Credit Scoring Systems with Quantum-Enhanced Machine Learning

Javier Mancilla, André Sequeira, Tomas Tagliani, Francisco Llaneza, Claudio Beiza

TL;DR

This work explores quantum-enhanced machine learning for credit scoring in FinTech, addressing data-scarce, imbalanced scenarios where classical models struggle. It introduces Systemic Quantum Score (SQS), an end-to-end evolutionary framework that discovers and refines quantum feature maps via Pauli-word encodings, guided by kernel target alignment. Through a FinTech case study (Fintonic) with a 350-feature risk dataset, SQS demonstrates superior performance over XGBoost and SVC in low-data regimes and shows robust generalization when data are scarce. The findings suggest quantum-kernel approaches can provide opportunistic gains in production-grade finance when data are limited or noisy, motivating further hardware-backed validation and broader domain testing.

Abstract

Quantum Kernels are projected to provide early-stage usefulness for quantum machine learning. However, highly sophisticated classical models are hard to surpass without losing interpretability, particularly when vast datasets can be exploited. Nonetheless, classical models struggle once data is scarce and skewed. Quantum feature spaces are projected to find better links between data features and the target class to be predicted even in such challenging scenarios and most importantly, enhanced generalization capabilities. In this work, we propose a novel approach called Systemic Quantum Score (SQS) and provide preliminary results indicating potential advantage over purely classical models in a production grade use case for the Finance sector. SQS shows in our specific study an increased capacity to extract patterns out of fewer data points as well as improved performance over data-hungry algorithms such as XGBoost, providing advantage in a competitive market as it is the FinTech and Neobank regime.

Empowering Credit Scoring Systems with Quantum-Enhanced Machine Learning

TL;DR

This work explores quantum-enhanced machine learning for credit scoring in FinTech, addressing data-scarce, imbalanced scenarios where classical models struggle. It introduces Systemic Quantum Score (SQS), an end-to-end evolutionary framework that discovers and refines quantum feature maps via Pauli-word encodings, guided by kernel target alignment. Through a FinTech case study (Fintonic) with a 350-feature risk dataset, SQS demonstrates superior performance over XGBoost and SVC in low-data regimes and shows robust generalization when data are scarce. The findings suggest quantum-kernel approaches can provide opportunistic gains in production-grade finance when data are limited or noisy, motivating further hardware-backed validation and broader domain testing.

Abstract

Quantum Kernels are projected to provide early-stage usefulness for quantum machine learning. However, highly sophisticated classical models are hard to surpass without losing interpretability, particularly when vast datasets can be exploited. Nonetheless, classical models struggle once data is scarce and skewed. Quantum feature spaces are projected to find better links between data features and the target class to be predicted even in such challenging scenarios and most importantly, enhanced generalization capabilities. In this work, we propose a novel approach called Systemic Quantum Score (SQS) and provide preliminary results indicating potential advantage over purely classical models in a production grade use case for the Finance sector. SQS shows in our specific study an increased capacity to extract patterns out of fewer data points as well as improved performance over data-hungry algorithms such as XGBoost, providing advantage in a competitive market as it is the FinTech and Neobank regime.
Paper Structure (13 sections, 6 equations, 6 figures, 3 tables, 1 algorithm)

This paper contains 13 sections, 6 equations, 6 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Background
  3. Quantum kernels
  4. Fintonic's challenge
  5. Loan application process
  6. Risk dataset
  7. Systemic Quantum Score (SQS)
  8. Evolutionary approach
  9. Results
  10. Obtained kernels
  11. Reduced datasets
  12. Generalization capacity
  13. Conclusions

Figures (6)

  • Figure 1: Quantum circuit illustrating the inversion test. $\phi(x)$ represents the quantum feature map. The probability of the all-zero is estimated as the kernel.
  • Figure 2: Basic examples of a translation between individual in the population and its circuit representation.
  • Figure 3: Basic visual representation of the evolutionary algorithm depicted in \ref{['alg:ea']}
  • Figure 4: Fidelity kernels obtained using 1000 individuals, 50 generations and 2 qubit for data mapping.
  • Figure 5: Fidelity kernels obtained using 1000 individuals, 50 generations and 5 qubit for data mapping for the 2000 down-sampled dataset.
  • ...and 1 more figures