Scalable Feature Selection for (Multitask) Gradient Boosted Trees

Cuize Han, Nikhil Rao, Daria Sorokina, Karthik Subbian
Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, PMLR 108:885-894, 2020.

Abstract

Gradient Boosted Decision Trees (GBDTs) are widely used for building ranking and relevance models in search and recommendation. Considerations such as latency and interpretability dictate the use of as few features as possible to train these models. Feature selection in GBDT models typically involves heuristically ranking the features by importance and selecting the top few, or by per- forming a full backward feature elimination routine. On-the-fly feature selection methods proposed previously scale suboptimally with the number of features, which can be daunt- ing in high dimensional settings. We develop a scalable forward feature selection variant for GBDT, via a novel group testing procedure that works well in high dimensions, and enjoys favorable theoretical performance and computational guarantees. We show via ex- tensive experiments on both public and proprietary datasets that the proposed method offers significant speedups in training time, while being as competitive as existing GBDT methods in terms of model performance metrics. We also extend the method to the multitask setting, allowing the practitioner to select common features across tasks, as well as selecting task-specific features.

Cite this Paper


BibTeX
@InProceedings{pmlr-v108-han20a, title = {Scalable Feature Selection for (Multitask) Gradient Boosted Trees}, author = {Han, Cuize and Rao, Nikhil and Sorokina, Daria and Subbian, Karthik}, booktitle = {Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics}, pages = {885--894}, year = {2020}, editor = {Chiappa, Silvia and Calandra, Roberto}, volume = {108}, series = {Proceedings of Machine Learning Research}, month = {26--28 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v108/han20a/han20a.pdf}, url = {http://proceedings.mlr.press/v108/han20a.html}, abstract = {Gradient Boosted Decision Trees (GBDTs) are widely used for building ranking and relevance models in search and recommendation. Considerations such as latency and interpretability dictate the use of as few features as possible to train these models. Feature selection in GBDT models typically involves heuristically ranking the features by importance and selecting the top few, or by per- forming a full backward feature elimination routine. On-the-fly feature selection methods proposed previously scale suboptimally with the number of features, which can be daunt- ing in high dimensional settings. We develop a scalable forward feature selection variant for GBDT, via a novel group testing procedure that works well in high dimensions, and enjoys favorable theoretical performance and computational guarantees. We show via ex- tensive experiments on both public and proprietary datasets that the proposed method offers significant speedups in training time, while being as competitive as existing GBDT methods in terms of model performance metrics. We also extend the method to the multitask setting, allowing the practitioner to select common features across tasks, as well as selecting task-specific features.} }
Endnote
%0 Conference Paper %T Scalable Feature Selection for (Multitask) Gradient Boosted Trees %A Cuize Han %A Nikhil Rao %A Daria Sorokina %A Karthik Subbian %B Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2020 %E Silvia Chiappa %E Roberto Calandra %F pmlr-v108-han20a %I PMLR %P 885--894 %U http://proceedings.mlr.press/v108/han20a.html %V 108 %X Gradient Boosted Decision Trees (GBDTs) are widely used for building ranking and relevance models in search and recommendation. Considerations such as latency and interpretability dictate the use of as few features as possible to train these models. Feature selection in GBDT models typically involves heuristically ranking the features by importance and selecting the top few, or by per- forming a full backward feature elimination routine. On-the-fly feature selection methods proposed previously scale suboptimally with the number of features, which can be daunt- ing in high dimensional settings. We develop a scalable forward feature selection variant for GBDT, via a novel group testing procedure that works well in high dimensions, and enjoys favorable theoretical performance and computational guarantees. We show via ex- tensive experiments on both public and proprietary datasets that the proposed method offers significant speedups in training time, while being as competitive as existing GBDT methods in terms of model performance metrics. We also extend the method to the multitask setting, allowing the practitioner to select common features across tasks, as well as selecting task-specific features.
APA
Han, C., Rao, N., Sorokina, D. & Subbian, K.. (2020). Scalable Feature Selection for (Multitask) Gradient Boosted Trees. Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 108:885-894 Available from http://proceedings.mlr.press/v108/han20a.html.

Related Material