Tabular data is the most commonly used type of data in industry, but deep learning on tabular data receives less attention than for CV and NLP. My aim in this talk is to review some of the popular DL models for Tabular data, explain their inner-works and compare results. After this talk you'll be familiar with the leading methods for modeling tabular data and able to apply it on your use case.
Throughout this talk we will review and explore the inner workings of a few of the following deep learning architectures for tabular data: SNN, NODE, TabNet, GrowNet, DCN V2, AutoInt, transformer for tabular data, simple MLP and ResNet. This will include: reviewing some of the architectures and the main advantages it offers for tabular data, and comparing performance on a regression problem over a public dataset (housing prices). This talk requires previous knowledge and is aimed at data scientists or others who are familiar with gradient boosting trees, basic deep learning i.e., you understand what a neural network is and how it works, and are familiar with activation functions and batch normalization.