Towards Fair In-Context Learning with Tabular Foundation Models

TL;DR

This paper investigates fairness in transformer-based tabular in-context learning models, proposing methods to reduce bias and demonstrating that uncertainty-based sample selection improves fairness with minimal accuracy loss.

Contribution

It is the first to analyze fairness in tabular ICL models and introduces three pre-processing methods to mitigate biases, with empirical validation on multiple datasets.

Findings

Uncertainty-based sample selection improves fairness metrics.

Fairness-enhancing methods have minimal impact on accuracy.

The study provides reproducible code for fairness in tabular ICL.

Abstract

Transformer-based tabular foundation models have recently demonstrated promising in-context learning (ICL) performance on structured data, emerging as competitive alternatives to gradient-boosted trees. However, the fairness implications of this new paradigm remain largely unexplored. We present the first investigation of fairness in tabular ICL, evaluating three recently proposed foundation models--TabPFNv2, TabICL, and TabDPT--on multiple benchmark datasets. To mitigate biases, we explore three pre-processing fairness-enhancing methods: correlation removal (decorrelating input features from the sensitive attribute), group-balanced sample selection (ensuring equal representation of protected groups in context examples), and uncertainty-based sample selection (prioritizing context examples with high sensitive-attribute prediction uncertainty). Our experiments show that the uncertainty-based strategy consistently improves group fairness metrics (e.g., demographic parity, equalized odds, and equal opportunity) with minimal impact on predictive accuracy. We release our code to facilitate reproducibility https://github.com/patrikken/Fair-TabICL.