Learning and privacy with incomplete data and weak supervision

Montreal, Canada on December 12th @ NIPS'15

Speakers: Wendy Cho, Kamalika Chaudhuri, Nando de Freitas, Max Ott
- with a special issue on the Journal of Privacy and Confidentiality -

Accepted contributions

In alphabetical order:

  • $(\varepsilon, \delta)$-differential privacy of Gibbs posteriors
    Kentaro Minami, Hiromi Arai and Issei Sato [pdf]

  • Alter-CNN: an approach for learning from label proportions with the application to ice-water Classification
    Fan Li and Graham Taylor [pdf]

  • Bridging weak supervision and privacy aware learning via sufficient statistics
    Giorgio Patrini, Frank Nielsen and Richard Nock [pdf]

  • DUAL-LOCO: Preserving privacy between features in distributed estimation
    Christina Heinze, Brian McWilliams and Nicolai Meinshausen [pdf]

  • Learning with differential privacy: stability, learnability and the sufficiency and necessity of ERM principle
    Yu-Xiang Wang, Jing Lei and Stephen E. Fienberg [pdf]

  • Message passing for collective graphical models
    Tao Sun, Daniel Sheldon and Akshat Kumar [pdf]

  • Private approximations of the 2nd-moment matrix using existing techniques in linear regression
    Or Sheffet [pdf]

  • Private posterior distributions from variational approximations
    Vishesh Karwa, Daniel Kifer and Aleksandra Slavkovic [pdf]