Matches in SemOpenAlex for { <https://semopenalex.org/work/W2950951261> ?p ?o ?g. }
- W2950951261 abstract "Although adversarial examples and model robustness have been extensively studied in the context of linear models and neural networks, research on this issue in tree-based models and how to make tree-based models robust against adversarial examples is still limited. In this paper, we show that tree based models are also vulnerable to adversarial examples and develop a novel algorithm to learn robust trees. At its core, our method aims to optimize the performance under the worst-case perturbation of input features, which leads to a max-min saddle point problem. Incorporating this saddle point objective into the decision tree building procedure is non-trivial due to the discrete nature of trees --- a naive approach to finding the best split according to this saddle point objective will take exponential time. To make our approach practical and scalable, we propose efficient tree building algorithms by approximating the inner minimizer in this saddle point problem, and present efficient implementations for classical information gain based trees as well as state-of-the-art tree boosting models such as XGBoost. Experimental results on real world datasets demonstrate that the proposed algorithms can substantially improve the robustness of tree-based models against adversarial examples." @default.
- W2950951261 created "2019-06-27" @default.
- W2950951261 creator A5010841999 @default.
- W2950951261 creator A5024120345 @default.
- W2950951261 creator A5062177930 @default.
- W2950951261 creator A5066048721 @default.
- W2950951261 date "2019-02-27" @default.
- W2950951261 modified "2023-09-25" @default.
- W2950951261 title "Robust Decision Trees Against Adversarial Examples" @default.
- W2950951261 cites W1594031697 @default.
- W2950951261 cites W1678356000 @default.
- W2950951261 cites W1987356990 @default.
- W2950951261 cites W2024046085 @default.
- W2950951261 cites W2070493638 @default.
- W2950951261 cites W2149706766 @default.
- W2950951261 cites W2153635508 @default.
- W2950951261 cites W2408141691 @default.
- W2950951261 cites W2592340788 @default.
- W2950951261 cites W2594717275 @default.
- W2950951261 cites W2607219512 @default.
- W2950951261 cites W2652074165 @default.
- W2950951261 cites W2740731087 @default.
- W2950951261 cites W2753111111 @default.
- W2950951261 cites W2765233338 @default.
- W2950951261 cites W2765384636 @default.
- W2950951261 cites W2768348081 @default.
- W2950951261 cites W2783692467 @default.
- W2950951261 cites W2786118190 @default.
- W2950951261 cites W2786163515 @default.
- W2950951261 cites W2787496614 @default.
- W2950951261 cites W2797359501 @default.
- W2950951261 cites W2898227265 @default.
- W2950951261 cites W2899692219 @default.
- W2950951261 cites W2908521812 @default.
- W2950951261 cites W2913266441 @default.
- W2950951261 cites W2942630857 @default.
- W2950951261 cites W2949103145 @default.
- W2950951261 cites W2962777143 @default.
- W2950951261 cites W2962943487 @default.
- W2950951261 cites W2962995403 @default.
- W2950951261 cites W2963143631 @default.
- W2950951261 cites W2963207607 @default.
- W2950951261 cites W2963389226 @default.
- W2950951261 cites W2963496101 @default.
- W2950951261 cites W2963539647 @default.
- W2950951261 cites W2963612069 @default.
- W2950951261 cites W2963793535 @default.
- W2950951261 cites W2963857521 @default.
- W2950951261 cites W2963879647 @default.
- W2950951261 cites W2964082701 @default.
- W2950951261 cites W2964116600 @default.
- W2950951261 cites W2964153729 @default.
- W2950951261 cites W2964253222 @default.
- W2950951261 cites W2964301649 @default.
- W2950951261 cites W2964346747 @default.
- W2950951261 cites W2979450790 @default.
- W2950951261 cites W2998277219 @default.
- W2950951261 doi "https://doi.org/10.48550/arxiv.1902.10660" @default.
- W2950951261 hasPublicationYear "2019" @default.
- W2950951261 type Work @default.
- W2950951261 sameAs 2950951261 @default.
- W2950951261 citedByCount "16" @default.
- W2950951261 countsByYear W29509512612019 @default.
- W2950951261 countsByYear W29509512612020 @default.
- W2950951261 countsByYear W29509512612021 @default.
- W2950951261 countsByYear W29509512612022 @default.
- W2950951261 countsByYear W29509512612023 @default.
- W2950951261 crossrefType "posted-content" @default.
- W2950951261 hasAuthorship W2950951261A5010841999 @default.
- W2950951261 hasAuthorship W2950951261A5024120345 @default.
- W2950951261 hasAuthorship W2950951261A5062177930 @default.
- W2950951261 hasAuthorship W2950951261A5066048721 @default.
- W2950951261 hasBestOaLocation W29509512611 @default.
- W2950951261 hasConcept C104317684 @default.
- W2950951261 hasConcept C113174947 @default.
- W2950951261 hasConcept C11413529 @default.
- W2950951261 hasConcept C119857082 @default.
- W2950951261 hasConcept C126255220 @default.
- W2950951261 hasConcept C134306372 @default.
- W2950951261 hasConcept C154945302 @default.
- W2950951261 hasConcept C185592680 @default.
- W2950951261 hasConcept C2524010 @default.
- W2950951261 hasConcept C2681867 @default.
- W2950951261 hasConcept C2777127463 @default.
- W2950951261 hasConcept C33923547 @default.
- W2950951261 hasConcept C37736160 @default.
- W2950951261 hasConcept C41008148 @default.
- W2950951261 hasConcept C46686674 @default.
- W2950951261 hasConcept C48044578 @default.
- W2950951261 hasConcept C55493867 @default.
- W2950951261 hasConcept C63479239 @default.
- W2950951261 hasConcept C77088390 @default.
- W2950951261 hasConcept C84525736 @default.
- W2950951261 hasConceptScore W2950951261C104317684 @default.
- W2950951261 hasConceptScore W2950951261C113174947 @default.
- W2950951261 hasConceptScore W2950951261C11413529 @default.
- W2950951261 hasConceptScore W2950951261C119857082 @default.
- W2950951261 hasConceptScore W2950951261C126255220 @default.
- W2950951261 hasConceptScore W2950951261C134306372 @default.
- W2950951261 hasConceptScore W2950951261C154945302 @default.