Citation
Bo Li, Yining Wang, Aarti Singh, Yevgeniy Vorobeychik. "Data poisoning attacks on factorization-based collaborative filtering". Neural Information Processing Systems (NIPS 2016), 2017.

Abstract
Recommendation and collaborative filtering systems are important in modern information and e-commerce applications. As these systems are becoming increasingly popular in the industry, their outputs could affect business decision making, introducing incentives for an adversarial party to compromise the availability or integrity of such systems. We introduce a data poisoning attack on collaborative filtering systems. We demonstrate how a powerful attacker with full knowledge of the learner can generate malicious data so as to maximize his/her malicious objectives, while at the same time mimicking normal user behavior to avoid being detected. While the complete knowledge assumption seems extreme, it enables a robust assessment of the vulnerability of collaborative filtering schemes to highly motivated attacks. We present efficient solutions for two popular factorizationbased collaborative filtering algorithms: the alternative minimization formulation and the nuclear norm minimization method. Finally, we test the effectiveness of our proposed algorithms on real-world data and discuss potential defensive strategies.

Electronic downloads

• http://vorobeychik.com/2016/collabpoison.pdf

• HTML

  Bo Li, Yining Wang, Aarti Singh, Yevgeniy Vorobeychik. <a
  href="http://www.cps-forces.org/pubs/252.html"
  >Data poisoning attacks on factorization-based
  collaborative filtering</a>, Neural Information
  Processing Systems (NIPS 2016), 2017.

• Plain text

  Bo Li, Yining Wang, Aarti Singh, Yevgeniy Vorobeychik.
  "Data poisoning attacks on factorization-based
  collaborative filtering". Neural Information Processing
  Systems (NIPS 2016), 2017.

• BibTeX

  @inproceedings{LiWangSinghVorobeychik17_DataPoisoningAttacksOnFactorizationbasedCollaborative,
      author = {Bo Li and Yining Wang and Aarti Singh and Yevgeniy
                Vorobeychik},
      title = {Data poisoning attacks on factorization-based
                collaborative filtering},
      booktitle = {Neural Information Processing Systems (NIPS 2016)},
      year = {2017},
      abstract = {Recommendation and collaborative filtering systems
                are important in modern information and e-commerce
                applications. As these systems are becoming
                increasingly popular in the industry, their
                outputs could affect business decision making,
                introducing incentives for an adversarial party to
                compromise the availability or integrity of such
                systems. We introduce a data poisoning attack on
                collaborative filtering systems. We demonstrate
                how a powerful attacker with full knowledge of the
                learner can generate malicious data so as to
                maximize his/her malicious objectives, while at
                the same time mimicking normal user behavior to
                avoid being detected. While the complete knowledge
                assumption seems extreme, it enables a robust
                assessment of the vulnerability of collaborative
                filtering schemes to highly motivated attacks. We
                present efficient solutions for two popular
                factorizationbased collaborative filtering
                algorithms: the alternative minimization
                formulation and the nuclear norm minimization
                method. Finally, we test the effectiveness of our
                proposed algorithms on real-world data and discuss
                potential defensive strategies. },
      URL = {http://cps-forces.org/pubs/252.html}
  }
  

Posted by Waseem Abbas on 2 Mar 2017.
Groups: forces
For additional information, see the Publications FAQ or contact webmaster at cps-forces org.