A framework for quantitative security analysis of machine learning
Authors:
Open Access
- 9 November 2009
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1-4
Abstract
We propose a framework for quantitative security analysis of machine learning methods. The key parts of this framework are the formal specification of a deployed learning model and attacker's constraints, the computation of an optimal attack, and the derivation of an upper bound on adversarial impact. We exemplarily apply the framework for the analysis of one specific learning scenario, online centroid anomaly detection, and experimentally verify the tightness of obtained theoretical boundsKeywords
This publication has 6 references indexed in Scilit:
- On the infeasibility of modeling polymorphic shellcode2007
- Evading network anomaly detection systems2006
- Can machine learning be secure?2006
- Misleading worm signature generators using deliberate noise injection2006
- Adversarial classification2004
- Learning in the Presence of Malicious ErrorsSIAM Journal on Computing, 1993