XBNet: An extremely boosted neural network

Abstract

• XBNet is a novel architecture and optimization technique for better performance, training stability and interpretability for tabular data, which is a trending research topic. • Easy to use with rapid prototyping capabilities • Opens the path to a new paradigm of neural networks combined with tree-based architectures which has very low code requirements and high performance. • Provides comprehensive details of the metrics with visualisation to facilitate the understanding of the training process and how it can be optimized. Neural networks have proved to be very robust at processing unstructured data like images, text, videos, and audio. However, it has been observed that their performance is not up to the mark in tabular data; hence tree-based models are preferred in such scenarios. A popular model for tabular data is boosted trees, a highly efficacious and extensively used machine learning method, and it also provides good interpretability compared to neural networks. In this paper, we describe a novel architecture XBNet (Extremely Boosted Neural Network), which tries to combine tree-based models with neural networks to create a robust architecture trained by using a novel optimization technique, Boosted Gradient Descent for Tabular Data which increases its interpretability and performance.