Manipulating Machine Learning: Poisoning Attacks and Countermeasures for Regression Learning

@article{Jagielski2018ManipulatingML,
  title={Manipulating Machine Learning: Poisoning Attacks and Countermeasures for Regression Learning},
  author={M. Jagielski and Alina Oprea and B. Biggio and C. Liu and C. Nita-Rotaru and Bo Li},
  journal={2018 IEEE Symposium on Security and Privacy (SP)},
  year={2018},
  pages={19-35}
}
As machine learning becomes widely used for automated decisions, attackers have strong incentives to manipulate the results and models generated by machine learning algorithms. [...] Key Method We propose a theoretically-grounded optimization framework specifically designed for linear regression and demonstrate its effectiveness on a range of datasets and models. We also introduce a fast statistical attack that requires limited knowledge of the training process.Expand
249 Citations
Data Poisoning Attacks on Regression Learning and Corresponding Defenses
  • Highly Influenced
  • PDF
Defending Regression Learners Against Poisoning Attacks
  • Highly Influenced
  • PDF
Poisoning Attacks with Generative Adversarial Nets
  • 16
  • PDF
Exploring Backdoor Poisoning Attacks Against Malware Classifiers
  • 10
  • PDF
A Flexible Poisoning Attack Against Machine Learning
  • 5
  • Highly Influenced
Subpopulation Data Poisoning Attacks
  • 10
  • PDF
On Embedding Backdoor in Malware Detectors Using Machine Learning
  • 1
Preventing Data Poisoning Attacks By Using Generative Models
  • 5
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 56 REFERENCES
ANTIDOTE: understanding and defending against poisoning of anomaly detectors
  • 278
  • Highly Influential
  • PDF
Evasion Attacks against Machine Learning at Test Time
  • 1,042
  • PDF
Systematic Poisoning Attacks on and Defenses for Machine Learning in Healthcare
  • 112
  • PDF
Membership Inference Attacks Against Machine Learning Models
  • 1,039
  • PDF
Using Machine Teaching to Identify Optimal Training-Set Attacks on Machine Learners
  • 235
  • Highly Influential
  • PDF
Poisoning Attacks against Support Vector Machines
  • 708
  • PDF
Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures
  • 852
  • PDF
Security analysis of online centroid anomaly detection
  • 68
  • Highly Influential
  • PDF
Practical Evasion of a Learning-Based Classifier: A Case Study
  • 287
  • PDF
...
1
2
3
4
5
...