Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures

@inproceedings{Fredrikson2015ModelIA,
  title={Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures},
  author={Matt Fredrikson and Somesh Jha and Thomas Ristenpart},
  booktitle={CCS '15},
  year={2015}
}
  • Matt Fredrikson, Somesh Jha, Thomas Ristenpart
  • Published in CCS '15 2015
  • Computer Science
  • Machine-learning (ML) algorithms are increasingly utilized in privacy-sensitive applications such as predicting lifestyle choices, making medical diagnoses, and facial recognition. In a model inversion attack, recently introduced in a case study of linear classifiers in personalized medicine by Fredrikson et al., adversarial access to an ML model is abused to learn sensitive genomic information about individuals. Whether model inversion attacks apply to settings outside theirs, however, is… CONTINUE READING

    Citations

    Publications citing this paper.
    SHOWING 1-10 OF 543 CITATIONS

    Systematic Evaluation of Privacy Risks of Machine Learning Models

    VIEW 1 EXCERPT
    CITES BACKGROUND

    Membership Model Inversion Attacks for Deep Networks

    VIEW 1 EXCERPT
    CITES BACKGROUND

    Membership Inference Attacks Against Machine Learning Models

    VIEW 4 EXCERPTS
    CITES METHODS & BACKGROUND
    HIGHLY INFLUENCED

    Stealing Machine Learning Models via Prediction APIs

    VIEW 1 EXCERPT
    CITES BACKGROUND

    Vulnerability Detection and Analysis in Adversarial Deep Learning

    The Unintended Consequences of Overfitting: Training Data Inference Attacks

    VIEW 7 EXCERPTS
    CITES BACKGROUND, METHODS & RESULTS

    FILTER CITATIONS BY YEAR

    2015
    2020

    CITATION STATISTICS

    • 56 Highly Influenced Citations

    • Averaged 161 Citations per year from 2018 through 2020

    References

    Publications referenced by this paper.

    The OpenCV library

    • G. Bradski
    • Dr. Dobb’s Journal of Software Tools,
    • 2000
    VIEW 3 EXCERPTS
    HIGHLY INFLUENTIAL