Defending against Poisoning Attacks in Online Learning Settings
Year: 2019
Type: conference-paper
Abstract: Machine learning systems are vulnerable to data poisoning, a coordinated attack where a fraction of the training dataset is manipulated by an attacker to subvert learning. In this paper we first formulate an optimal attack strategy against online learning classifiers to assess worst-case scenarios. We also propose two defence mechanisms to mitigate the effect of online poisoning attacks by analysing the impact of the data points in the classifier and by means of an adaptive combination of machine learning classifiers with different learning rates. Our experimental evaluation supports the usefulness of our proposed defences to mitigate the effect of poisoning attacks in online learning settings.
Source: Spiral (Imperial College London)
Authors: Greg Collinge, Emil Lupu, Luis Muñoz-González
Institution: Imperial College London
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: European Commission
OpenAlex ID: W3013223614
