Born-Again Tree Ensembles

Thibaut Vidal, Maximilian Schiffer
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:9743-9753, 2020.

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, in particular, 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.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-vidal20a, title = {Born-Again Tree Ensembles}, author = {Vidal, Thibaut and Schiffer, Maximilian}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {9743--9753}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/vidal20a/vidal20a.pdf}, url = {https://proceedings.mlr.press/v119/vidal20a.html}, 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, in particular, 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.} }
Endnote
%0 Conference Paper %T Born-Again Tree Ensembles %A Thibaut Vidal %A Maximilian Schiffer %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-vidal20a %I PMLR %P 9743--9753 %U https://proceedings.mlr.press/v119/vidal20a.html %V 119 %X 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, in particular, 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.
APA
Vidal, T. & Schiffer, M.. (2020). Born-Again Tree Ensembles. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:9743-9753 Available from https://proceedings.mlr.press/v119/vidal20a.html.

Related Material