Data Poisoning Attacks against Online Learning
arXiv:1808.08994
Abstract
We consider data poisoning attacks, a class of adversarial attacks on machine learning where an adversary has the power to alter a small fraction of the training data in order to make the trained classifier satisfy certain objectives. While there has been much prior work on data poisoning, most of it is in the offline setting, and attacks for online learning, where training data arrives in a streaming manner, are not well understood. In this work, we initiate a systematic investigation of data poisoning attacks for online learning. We formalize the problem into two settings, and we propose a general attack strategy, formulated as an optimization problem, that applies to both with some modifications. We propose three solution strategies, and perform extensive experimental evaluation. Finally, we discuss the implications of our findings for building successful defenses.
Cited by in corpus (19)
- Optimization Problems for Machine Learning: A Survey
- Adversarial Machine Learning in Image Classification: A Survey Towards the Defender's Perspective
- TrISec: Training Data-Unaware Imperceptible Security Attacks on Deep Neural Networks
- Online Data Poisoning Attack
- Vulnerability-Aware Poisoning Mechanism for Online RL with Unknown Dynamics
- Towards Security Threats of Deep Learning Systems: A Survey
- FooBaR: Fault Fooling Backdoor Attack on Neural Network Training
- Model-Targeted Poisoning Attacks with Provable Convergence
- DeepObliviate: A Powerful Charm for Erasing Data Residual Memory in Deep Neural Networks
- Using Machine Teaching to Investigate Human Assumptions when Teaching Reinforcement Learners
- An Investigation of Data Poisoning Defenses for Online Learning
- Gradient-based Data Subversion Attack Against Binary Classifiers
- The Sample Complexity of Teaching-by-Reinforcement on Q-Learning
- A Separation Result Between Data-oblivious and Data-aware Poisoning Attacks
- Extending Adversarial Attacks to Produce Adversarial Class Probability Distributions
- Accumulative Poisoning Attacks on Real-time Data
- Lethean Attack: An Online Data Poisoning Technique
- FAdeML: Understanding the Impact of Pre-Processing Noise Filtering on Adversarial Machine Learning
- Shapley Homology: Topological Analysis of Sample Influence for Neural Networks