Corpus ID: 220381492

Oracle Efficient Private Non-Convex Optimization

  title={Oracle Efficient Private Non-Convex Optimization},
  author={Seth Neel and A. Roth and G. Vietri and Z. Wu},
  booktitle={ICML 2020},
  • Seth Neel, A. Roth, +1 author Z. Wu
  • Published in ICML 2020
  • Computer Science, Mathematics
  • One of the most effective algorithms for differentially private learning and optimization is objective perturbation. This technique augments a given optimization problem (e.g. deriving from an ERM problem) with a random linear term, and then exactly solves it. However prior analyses of this approach crucially rely on the convexity and smoothness of the objective function. We give two algorithms that extend this approach substantially. The first algorithm requires nothing except boundedness of… CONTINUE READING
    1 Citations
    New Oracle-Efficient Algorithms for Private Synthetic Data Release
    • 3
    • PDF


    Private Empirical Risk Minimization: Efficient Algorithms and Tight Error Bounds
    • 343
    • PDF
    Private Convex Empirical Risk Minimization and High-dimensional Regression
    • 175
    • PDF
    Towards Practical Differentially Private Convex Optimization
    • 35
    • PDF
    How to Use Heuristics for Differential Privacy
    • Seth Neel, A. Roth, Z. Wu
    • Computer Science, Mathematics
    • 2019 IEEE 60th Annual Symposium on Foundations of Computer Science (FOCS)
    • 2019
    • 16
    • PDF
    Differentially Private Empirical Risk Minimization
    • 757
    • Highly Influential
    • PDF
    Differentially private combinatorial optimization
    • 119
    • PDF
    A learning theory approach to noninteractive database privacy
    • 247
    • PDF
    The Algorithmic Foundations of Differential Privacy
    • 2,324
    • PDF
    Efficient Algorithms for Adversarial Contextual Learning
    • 48
    • PDF
    Mechanism Design via Differential Privacy
    • F. McSherry, Kunal Talwar
    • Computer Science
    • 48th Annual IEEE Symposium on Foundations of Computer Science (FOCS'07)
    • 2007
    • 1,484
    • PDF