Born-Again Tree Ensembles

Abstract

The use of machine learning algorithms in finance, medicine, and criminal justice can deeply impact human lives. As a consequence, research into interpretable machine learning has rapidly grown in an attempt to better control and fix possible sources of mistakes and biases. Tree ensembles offer a good prediction quality in various domains, but the concurrent use of multiple trees reduces the interpretability of the ensemble. Against this background, we study born-again tree ensembles, i.e., the process of constructing a single decision tree of minimum size that reproduces the exact same behavior as a given tree ensemble in its entire feature space. To find such a tree, we develop a dynamic-programming based algorithm that exploits sophisticated pruning and bounding rules to reduce the number of recursive calls. This algorithm generates optimal born-again trees for many datasets of practical interest, leading to classifiers which are typically simpler and more interpretable without any other form of compromise.

arXiv comment"Born-Again Tree Ensembles", proceedings of ICML 2020. The associated source code is available at: https://github.com/vidalt/BA-Trees

References in corpus (16)

• Distilling the Knowledge in a Neural Network
• Interpretable machine learning: definitions, methods, and applications
• Predictive learning via rule ensembles
• The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
• Born Again Neural Networks
• Towards Robust Interpretability with Self-Explaining Neural Networks
• Distilling a Neural Network Into a Soft Decision Tree
• Interpreting Blackbox Models via Model Extraction
• Node harvest
• Interpretability via Model Extraction
• Interpreting Models via Single Tree Approximation
• Tree Space Prototypes: Another Look at Making Tree Ensembles Interpretable
• Optimal Generalized Decision Trees via Integer Programming
• Making Tree Ensembles Interpretable: A Bayesian Model Selection Approach
• Rectified Decision Trees: Towards Interpretability, Compression and Empirical Soundness
• LionForests: Local Interpretation of Random Forests

Cited by in corpus (8)

• Security and privacy for 6G: A survey on prospective technologies and challenges
• FS-BAN: Born-Again Networks for Domain Generalization Few-Shot Classification
• MurTree: Optimal Classification Trees via Dynamic Programming and Search
• Fast, Accurate, and Simple Models for Tabular Data via Augmented Distillation
• Provably efficient, succinct, and precise explanations
• Optimal Counterfactual Explanations in Tree Ensembles
• On the price of explainability for some clustering problems
• Making CNNs Interpretable by Building Dynamic Sequential Decision Forests with Top-down Hierarchy Learning