Adversarial training is a prominent approach to make machine learning (ML) models resilient to adversarial examples. Unfortunately, such approach assumes the use of differentiable learning models, hence it cannot be applied to relevant ML techniques, such as ensembles of decision trees. In this paper, we generalize adversarial training to gradient-boosted decision trees (GBDTs). Our experiments show that the performance of classifiers based on existing learning techniques either sharply decreases upon attack or is unsatisfactory in absence of attacks, while adversarial training provides a very good trade-off between resiliency to attacks and accuracy in the unattacked setting.

Adversarial training of gradient-boosted decision trees

Calzavara S.;Lucchese C.;
2019-01-01

Abstract

Adversarial training is a prominent approach to make machine learning (ML) models resilient to adversarial examples. Unfortunately, such approach assumes the use of differentiable learning models, hence it cannot be applied to relevant ML techniques, such as ensembles of decision trees. In this paper, we generalize adversarial training to gradient-boosted decision trees (GBDTs). Our experiments show that the performance of classifiers based on existing learning techniques either sharply decreases upon attack or is unsatisfactory in absence of attacks, while adversarial training provides a very good trade-off between resiliency to attacks and accuracy in the unattacked setting.
2019
International Conference on Information and Knowledge Management, Proceedings
File in questo prodotto:
File Dimensione Formato  
cikm19.pdf

non disponibili

Tipologia: Documento in Post-print
Licenza: Accesso chiuso-personale
Dimensione 725.17 kB
Formato Adobe PDF
725.17 kB Adobe PDF   Visualizza/Apri

I documenti in ARCA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10278/3722896
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 24
  • ???jsp.display-item.citation.isi??? 21
social impact