DANets: Deep Abstract Networks for Tabular Data Classification and Regression

TL;DR

DANets introduces a novel neural network architecture with Abstract Layers for improved classification and regression on tabular data, emphasizing feature grouping, semantic abstraction, and computational efficiency.

Contribution

The paper presents a new neural component called Abstract Layer and a deep network structure, DANets, tailored specifically for tabular data, with a re-parameterization method for efficiency.

Findings

Effective on seven real-world datasets.

Outperforms competitive methods in accuracy and efficiency.

Demonstrates good extendibility with increased depth.

Abstract

Tabular data are ubiquitous in real world applications. Although many commonly-used neural components (e.g., convolution) and extensible neural networks (e.g., ResNet) have been developed by the machine learning community, few of them were effective for tabular data and few designs were adequately tailored for tabular data structures. In this paper, we propose a novel and flexible neural component for tabular data, called Abstract Layer (AbstLay), which learns to explicitly group correlative input features and generate higher-level features for semantics abstraction. Also, we design a structure re-parameterization method to compress the learned AbstLay, thus reducing the computational complexity by a clear margin in the reference phase. A special basic block is built using AbstLays, and we construct a family of Deep Abstract Networks (DANets) for tabular data classification and regression by stacking such blocks. In DANets, a special shortcut path is introduced to fetch information from raw tabular features, assisting feature interactions across different levels. Comprehensive experiments on seven real-world tabular datasets show that our AbstLay and DANets are effective for tabular data classification and regression, and the computational complexity is superior to competitive methods. Besides, we evaluate the performance gains of DANet as it goes deep, verifying the extendibility of our method. Our code is available at https://github.com/WhatAShot/DANet.