Distribution Shift Aware Neural Tabular Learning

TL;DR

This paper introduces SAFT, a novel framework for improving neural tabular learning robustness under distribution shifts by transforming features into a continuous, optimizable space, leading to better generalization.

Contribution

The paper formalizes the DSTL problem and proposes SAFT, a shift-aware feature transformation method that enhances robustness and generalization in tabular learning under distribution shifts.

Findings

SAFT outperforms prior methods under diverse distribution shifts.

SAFT improves robustness and generalization in real-world scenarios.

Extensive experiments validate the effectiveness of SAFT.

Abstract

Tabular learning transforms raw features into optimized spaces for downstream tasks, but its effectiveness deteriorates under distribution shifts between training and testing data. We formalize this challenge as the Distribution Shift Tabular Learning (DSTL) problem and propose a novel Shift-Aware Feature Transformation (SAFT) framework to address it. SAFT reframes tabular learning from a discrete search task into a continuous representation-generation paradigm, enabling differentiable optimization over transformed feature sets. SAFT integrates three mechanisms to ensure robustness: (i) shift-resistant representation via embedding decorrelation and sample reweighting, (ii) flatness-aware generation through suboptimal embedding averaging, and (iii) normalization-based alignment between training and test distributions. Extensive experiments show that SAFT consistently outperforms prior tabular learning methods in terms of robustness, effectiveness, and generalization ability under diverse real-world distribution shifts.