ExactBoost: Directly Boosting the Margin in Combinatorial and Non-decomposable Metrics

Daniel Csillag, Carolina Piazza, Thiago Ramos, João Vitor Romano, Roberto I. Oliveira, Paulo Orenstein
Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, PMLR 151:9017-9049, 2022.

Abstract

Many classification algorithms require the use of surrogate losses when the intended loss function is combinatorial or non-decomposable. This paper introduces a fast and exact stagewise optimization algorithm, dubbed ExactBoost, that boosts stumps to the actual loss function. By developing a novel extension of margin theory to the non-decomposable setting, it is possible to provably bound the generalization error of ExactBoost for many important metrics with different levels of non-decomposability. Through extensive examples, it is shown that such theoretical guarantees translate to competitive empirical performance. In particular, when used as an ensembler, ExactBoost is able to significantly outperform other surrogate-based and exact algorithms available.

Cite this Paper

BibTeX
@InProceedings{pmlr-v151-csillag22a, title = { ExactBoost: Directly Boosting the Margin in Combinatorial and Non-decomposable Metrics }, author = {Csillag, Daniel and Piazza, Carolina and Ramos, Thiago and Vitor Romano, Jo\~ao and Oliveira, Roberto I. and Orenstein, Paulo}, booktitle = {Proceedings of The 25th International Conference on Artificial Intelligence and Statistics}, pages = {9017--9049}, year = {2022}, editor = {Camps-Valls, Gustau and Ruiz, Francisco J. R. and Valera, Isabel}, volume = {151}, series = {Proceedings of Machine Learning Research}, month = {28--30 Mar}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v151/csillag22a/csillag22a.pdf}, url = {https://proceedings.mlr.press/v151/csillag22a.html}, abstract = { Many classification algorithms require the use of surrogate losses when the intended loss function is combinatorial or non-decomposable. This paper introduces a fast and exact stagewise optimization algorithm, dubbed ExactBoost, that boosts stumps to the actual loss function. By developing a novel extension of margin theory to the non-decomposable setting, it is possible to provably bound the generalization error of ExactBoost for many important metrics with different levels of non-decomposability. Through extensive examples, it is shown that such theoretical guarantees translate to competitive empirical performance. In particular, when used as an ensembler, ExactBoost is able to significantly outperform other surrogate-based and exact algorithms available. } }
Endnote
%0 Conference Paper %T ExactBoost: Directly Boosting the Margin in Combinatorial and Non-decomposable Metrics %A Daniel Csillag %A Carolina Piazza %A Thiago Ramos %A João Vitor Romano %A Roberto I. Oliveira %A Paulo Orenstein %B Proceedings of The 25th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2022 %E Gustau Camps-Valls %E Francisco J. R. Ruiz %E Isabel Valera %F pmlr-v151-csillag22a %I PMLR %P 9017--9049 %U https://proceedings.mlr.press/v151/csillag22a.html %V 151 %X Many classification algorithms require the use of surrogate losses when the intended loss function is combinatorial or non-decomposable. This paper introduces a fast and exact stagewise optimization algorithm, dubbed ExactBoost, that boosts stumps to the actual loss function. By developing a novel extension of margin theory to the non-decomposable setting, it is possible to provably bound the generalization error of ExactBoost for many important metrics with different levels of non-decomposability. Through extensive examples, it is shown that such theoretical guarantees translate to competitive empirical performance. In particular, when used as an ensembler, ExactBoost is able to significantly outperform other surrogate-based and exact algorithms available.
APA
Csillag, D., Piazza, C., Ramos, T., Vitor Romano, J., Oliveira, R.I. & Orenstein, P.. (2022). ExactBoost: Directly Boosting the Margin in Combinatorial and Non-decomposable Metrics . Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 151:9017-9049 Available from https://proceedings.mlr.press/v151/csillag22a.html.

Related Material