Table Transformers for Imputing Textual Attributes

TL;DR

This paper introduces Table Transformers for Imputing Textual Attributes (TTITA), an end-to-end transformer-based method that effectively imputes missing textual data in tables, outperforming traditional models especially on longer sequences.

Contribution

The paper presents a novel transformer-based approach for imputing missing textual data in tabular datasets, incorporating multi-task learning for heterogeneous columns and demonstrating superior performance.

Findings

Outperforms baseline models like RNNs and Llama2.

Significant improvements with longer target sequences.

Multi-task learning enhances imputation accuracy.

Abstract

Missing data in tabular dataset is a common issue as the performance of downstream tasks usually depends on the completeness of the training dataset. Previous missing data imputation methods focus on numeric and categorical columns, but we propose a novel end-to-end approach called Table Transformers for Imputing Textual Attributes (TTITA) based on the transformer to impute unstructured textual columns using other columns in the table. We conduct extensive experiments on three datasets, and our approach shows competitive performance outperforming baseline models such as recurrent neural networks and Llama2. The performance improvement is more significant when the target sequence has a longer length. Additionally, we incorporate multi-task learning to simultaneously impute for heterogeneous columns, boosting the performance for text imputation. We also qualitatively compare with ChatGPT for realistic applications.