Making Pre-trained Language Models Great on Tabular Prediction

TL;DR

This paper introduces TP-BERTa, a pre-trained language model tailored for tabular data prediction, using a novel tokenization method to convert numerical features into discrete tokens, improving transferability and performance.

Contribution

The paper proposes TP-BERTa, a novel pre-trained language model for tabular data that employs relative magnitude tokenization and intra-feature attention to enhance prediction accuracy.

Findings

TP-BERTa outperforms existing tabular DNNs in experiments.

TP-BERTa is competitive with Gradient Boosted Decision Trees.

The model effectively integrates feature names and values for better predictions.

Abstract

The transferability of deep neural networks (DNNs) has made significant progress in image and language processing. However, due to the heterogeneity among tables, such DNN bonus is still far from being well exploited on tabular data prediction (e.g., regression or classification tasks). Condensing knowledge from diverse domains, language models (LMs) possess the capability to comprehend feature names from various tables, potentially serving as versatile learners in transferring knowledge across distinct tables and diverse prediction tasks, but their discrete text representation space is inherently incompatible with numerical feature values in tables. In this paper, we present TP-BERTa, a specifically pre-trained LM for tabular data prediction. Concretely, a novel relative magnitude tokenization converts scalar numerical feature values to finely discrete, high-dimensional tokens, and an intra-feature attention approach integrates feature values with the corresponding feature names. Comprehensive experiments demonstrate that our pre-trained TP-BERTa leads the performance among tabular DNNs and is competitive with Gradient Boosted Decision Tree models in typical tabular data regime.