Team for Research in
Ubiquitous Secure Technology

Citation
"Open Problems in the Security of Learning.". M. Barreno, P. Bartlett, F. Chi, A. Joseph, B. Nelson, B. Rubinstein, U. Saini, and J. D. Tygar (eds.), Proceedings of the First ACM Workshop on AISec, 2008.

Abstract
Machine learning has become a valuable tool for detecting and preventing malicious activity. However, as more ap- plications employ machine learning techniques in adversar- ial decision-making situations, increasingly powerful attacks become possible against machine learning systems. In this paper, we present three broad research directions towards the end of developing truly secure learning. First, we suggest that nding bounds on adversarial in uence is important to understand the limits of what an attacker can and cannot do to a learning system. Second, we investigate the value of adversarial capabilities|the success of an attack depends largely on what types of information and in uence the at- tacker has. Finally, we propose directions in technologies for secure learning and suggest lines of investigation into secure techniques for learning in adversarial environments. We intend this paper to foster discussion about the security of machine learning, and we believe that the research direc- tions we propose represent the most important directions to pursue in the quest for secure learning.

Electronic downloads

• SML_open_problems.pdf · application/pdf · 378 kbytes

• HTML

   <a
  href="http://www.truststc.org/pubs/744.html"
  ><i>Open Problems in the Security of
  Learning.</i></a>,  M. Barreno, P. Bartlett, F.
  Chi, A. Joseph, B. Nelson, B. Rubinstein, U. Saini, and J.
  D. Tygar (eds.), Proceedings of the First ACM Workshop on
  AISec, 2008.

• Plain text

   "Open Problems in the Security of Learning.".  M.
  Barreno, P. Bartlett, F. Chi, A. Joseph, B. Nelson, B.
  Rubinstein, U. Saini, and J. D. Tygar (eds.), Proceedings of
  the First ACM Workshop on AISec, 2008.

• BibTeX

  @proceedings{BarrenoBartlettChiJosephNelsonRubinsteinSainiTygar08_OpenProblemsInSecurityOfLearning,
      title = {Open Problems in the Security of Learning.},
      editor = { M. Barreno, P. Bartlett, F. Chi, A. Joseph, B.
                Nelson, B. Rubinstein, U. Saini, and J. D. Tygar},
      organization = {Proceedings of the First ACM Workshop on AISec},
      year = {2008},
      abstract = {Machine learning has become a valuable tool for
                detecting and preventing malicious activity.
                However, as more ap- plications employ machine
                learning techniques in adversar- ial
                decision-making situations, increasingly powerful
                attacks become possible against machine learning
                systems. In this paper, we present three broad
                research directions towards the end of developing
                truly secure learning. First, we suggest that
                nding bounds on adversarial in uence is important
                to understand the limits of what an attacker can
                and cannot do to a learning system. Second, we
                investigate the value of adversarial
                capabilities|the success of an attack depends
                largely on what types of information and in uence
                the at- tacker has. Finally, we propose directions
                in technologies for secure learning and suggest
                lines of investigation into secure techniques for
                learning in adversarial environments. We intend
                this paper to foster discussion about the security
                of machine learning, and we believe that the
                research direc- tions we propose represent the
                most important directions to pursue in the quest
                for secure learning.},
      URL = {http://www.truststc.org/pubs/744.html}
  }
  

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