AI & Computational Science

AI Models Revolutionize How Banks Predict Credit Risk and Defaults

How the science connects

Machine learningPredictive modelingCredit risk

AI Insight

This study benchmarks tabular foundation models against traditional machine learning methods for credit risk prediction, specifically for Probability of Default and Loss Given Default modeling. Testing 33 classification methods across 14 datasets and 27 regression methods across 7 datasets, the research found that recent tabular foundation models (TabPFN-3 and TabICLv2) achieved superior average performance compared to conventional approaches like gradient-boosting models, particularly in scenarios with small sample sizes and highly imbalanced datasets. Notably, these foundation models operated without task-specific training or hyperparameter optimization, using only their pretrained parameters.


The findings suggest that tabular foundation models could significantly improve credit risk assessment in challenging real-world scenarios, such as low-default portfolios where traditional models struggle due to limited data. This advancement could enhance lending decision-making accuracy and financial system stability while reducing the need for extensive model tuning and training on institution-specific data.


⚠️ Preprint – Noch nicht peer-reviewed

Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.

Abstract: Predictive models are central to credit risk management, where the accuracy of default-probability and loss estimates affects the profitability of lending and the stability of the financial system. Decades of benchmarking research have consolidated the state of the art, with gradient-boosting models often representing the strongest machine-learning benchmark, while linear and logistic models remain important regulatory and industry reference models. The recent advent of tabular data foundation models introduces a new paradigm for predictive modeling. Rather than fitting a model to a target dataset, they use in-context learning to integrate information from the target dataset with prior knowledge acquired through large-scale pretraining on diverse collections of (synthetic) tabular data. We conjecture that access to information beyond the target data is beneficial in small-data settings and may help address longstanding challenges, including low-default portfolios. Whether this promise holds in practice remains an open question. This paper benchmarks state-of-the-art tabular foundation models against a broad set of classical and advanced machine learning competitors on two core tasks: Probability of Default (PD) and Loss Given Default (LGD) modeling. We evaluate $33$ classification methods across $14$ PD datasets and $27$ regression methods across $7$ LGD datasets, spanning a range of performance indicators and experimental conditions. The most recent tabular foundation models, in particular TabPFN-3 and TabICLv2, achieve the strongest average results, with the largest gains in small-sample and strongly imbalanced settings. These results are notable because these models are used with their pretrained parameters fixed and without downstream hyperparameter optimization or task-specific parameter updates.

Source: Tabular Foundation Models for Credit Risk Prediction: A Dedicated Review and Empirical Comparison