TabPFN Unleashed: A Scalable and Effective Solution to Tabular Classification Problems

TL;DR

This paper introduces Beta, a novel method that enhances TabPFN's performance on tabular classification by reducing bias and variance, improving scalability, and maintaining efficiency, validated on over 200 datasets.

Contribution

Beta is a new approach that combines lightweight encoding, bagging, and fine-tuning to improve TabPFN's adaptability and performance on diverse, large-scale tabular data.

Findings

Beta outperforms or matches state-of-the-art on 200+ datasets.

The method effectively handles high-dimensional data.

It maintains computational efficiency during inference.

Abstract

TabPFN has emerged as a promising in-context learning model for tabular data, capable of directly predicting the labels of test samples given labeled training examples. It has demonstrated competitive performance, particularly on small-scale classification tasks. However, despite its effectiveness, TabPFN still requires further refinement in several areas, including handling high-dimensional features, aligning with downstream datasets, and scaling to larger datasets. In this paper, we revisit existing variants of TabPFN and observe that most approaches focus either on reducing bias or variance, often neglecting the need to address the other side, while also increasing inference overhead. To fill this gap, we propose Beta (Bagging and Encoder-based Fine-tuning for TabPFN Adaptation), a novel and effective method designed to minimize both bias and variance. To reduce bias, we introduce a lightweight encoder to better align downstream tasks with the pre-trained TabPFN. By increasing the number of encoders in a lightweight manner, Beta mitigate variance, thereby further improving the model's performance. Additionally, bootstrapped sampling is employed to further reduce the impact of data perturbations on the model, all while maintaining computational efficiency during inference. Our approach enhances TabPFN's ability to handle high-dimensional data and scale to larger datasets. Experimental results on over 200 benchmark classification datasets demonstrate that Beta either outperforms or matches state-of-the-art methods.