NBDT: Neural-Backed Decision Trees

TL;DR

NBDTs integrate decision trees with neural networks to enhance both accuracy and interpretability, achieving competitive results on image datasets while providing clearer insights into model decisions.

Contribution

This work introduces Neural-Backed Decision Trees, a novel approach that jointly improves accuracy and interpretability by replacing neural network layers with differentiable decision sequences.

Findings

NBDTs match or outperform modern neural networks on CIFAR and ImageNet.

NBDTs generalize to unseen classes by up to 16%.

Surrogate loss improves original model accuracy by up to 2%.

Abstract

Machine learning applications such as finance and medicine demand accurate and justifiable predictions, barring most deep learning methods from use. In response, previous work combines decision trees with deep learning, yielding models that (1) sacrifice interpretability for accuracy or (2) sacrifice accuracy for interpretability. We forgo this dilemma by jointly improving accuracy and interpretability using Neural-Backed Decision Trees (NBDTs). NBDTs replace a neural network's final linear layer with a differentiable sequence of decisions and a surrogate loss. This forces the model to learn high-level concepts and lessens reliance on highly-uncertain decisions, yielding (1) accuracy: NBDTs match or outperform modern neural networks on CIFAR, ImageNet and better generalize to unseen classes by up to 16%. Furthermore, our surrogate loss improves the original model's accuracy by up to 2%. NBDTs also afford (2) interpretability: improving human trustby clearly identifying model mistakes and assisting in dataset debugging. Code and pretrained NBDTs are at https://github.com/alvinwan/neural-backed-decision-trees.