Exploiting machine learning to subvert your spam filter

Citation
"Exploiting machine learning to subvert your spam filter". B. Nelson, M. Barreno, F. Chi, A. D. Joseph, B. Rubinstein, U. Saini, C. Sutton, J. D. Tygar, and Kai Xia (eds.), Proceedings of the First USENIX Workshop on Large-Scale Exploits and Emergent Threats, April, 2008.

Abstract
Using statistical machine learning for making security decisions introduces new vulnerabilities in large scale systems. This paper shows how an adversary can exploit statistical machine learning, as used in the SpamBayes spam filter, to render it useless—even if the adversary’s access is limited to only 1% of the training messages. We further demonstrate a new class of focused attacks that successfully prevent victims from receiving specific email messages. Finally, we introduce two new types of defenses against these attacks.

Electronic downloads

• Spam_filter.pdf · application/pdf · 419 kbytes

Citation formats

• HTML

   <a
  href="http://www.truststc.org/pubs/747.html"
  ><i>Exploiting machine learning to subvert your
  spam filter</i></a>,  B. Nelson, M. Barreno, F.
  Chi, A. D. Joseph, B. Rubinstein, U. Saini, C. Sutton, J. D.
  Tygar, and Kai Xia (eds.),  Proceedings of the First USENIX
  Workshop on Large-Scale Exploits and Emergent Threats,
  April, 2008.

• Plain text

   "Exploiting machine learning to subvert your spam
  filter".  B. Nelson, M. Barreno, F. Chi, A. D. Joseph,
  B. Rubinstein, U. Saini, C. Sutton, J. D. Tygar, and Kai Xia
  (eds.),  Proceedings of the First USENIX Workshop on
  Large-Scale Exploits and Emergent Threats, April, 2008.

• BibTeX

  @proceedings{NelsonBarrenoChiJosephRubinsteinSainiSuttonTygarXia08_ExploitingMachineLearningToSubvertYourSpamFilter,
      title = {Exploiting machine learning to subvert your spam
                filter},
      editor = { B. Nelson, M. Barreno, F. Chi, A. D. Joseph, B.
                Rubinstein, U. Saini, C. Sutton, J. D. Tygar, and
                Kai Xia},
      organization = { Proceedings of the First USENIX Workshop on
                Large-Scale Exploits and Emergent Threats},
      month = {April},
      year = {2008},
      abstract = {Using statistical machine learning for making
                security decisions introduces new vulnerabilities
                in large scale systems. This paper shows how an
                adversary can exploit statistical machine
                learning, as used in the SpamBayes spam filter, to
                render it useless—even if the adversary’s
                access is limited to only 1% of the training
                messages. We further demonstrate a new class of
                focused attacks that successfully prevent victims
                from receiving specific email messages. Finally,
                we introduce two new types of defenses against
                these attacks.},
      URL = {http://www.truststc.org/pubs/747.html}
  }
  

Posted by Jessica Gamble on 4 May 2010.
For additional information, see the Publications FAQ or contact webmaster at www truststc org..