TabINR: An Implicit Neural Representation Framework for Tabular Data Imputation
This work addresses the need for robust and high-quality imputation methods to improve downstream model performance in applications relying on tabular data, representing an incremental advance over prior deep learning approaches.
The paper tackles the problem of missing values in tabular data by introducing TabINR, an implicit neural representation framework for imputation, which demonstrates strong accuracy across twelve real-world datasets, often matching or outperforming existing methods, with notable gains on high-dimensional datasets.
Tabular data builds the basis for a wide range of applications, yet real-world datasets are frequently incomplete due to collection errors, privacy restrictions, or sensor failures. As missing values degrade the performance or hinder the applicability of downstream models, and while simple imputing strategies tend to introduce bias or distort the underlying data distribution, we require imputers that provide high-quality imputations, are robust across dataset sizes and yield fast inference. We therefore introduce TabINR, an auto-decoder based Implicit Neural Representation (INR) framework that models tables as neural functions. Building on recent advances in generalizable INRs, we introduce learnable row and feature embeddings that effectively deal with the discrete structure of tabular data and can be inferred from partial observations, enabling instance adaptive imputations without modifying the trained model. We evaluate our framework across a diverse range of twelve real-world datasets and multiple missingness mechanisms, demonstrating consistently strong imputation accuracy, mostly matching or outperforming classical (KNN, MICE, MissForest) and deep learning based models (GAIN, ReMasker), with the clearest gains on high-dimensional datasets.