Adversarial concept drift detection under poisoning attacks for robust data stream mining

DOI10.1007/S10994-022-06177-WMaRDI QIDQ6053814zbMATH OpenOpenAlexWikidataFDO

Authors Łukasz Korycki, Bartosz Krawczyk

Publication date 24 October 2023

Published in Machine Learning (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2009.09497

zbMATH Keywords

adversarial learningBoltzmann machineconcept driftdata stream miningpoisoning attacksrobust machine learning

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Online algorithms; streaming algorithms (68W27)

Abstract: Continuous learning from streaming data is among the most challenging topics in the contemporary machine learning. In this domain, learning algorithms must not only be able to handle massive volumes of rapidly arriving data, but also adapt themselves to potential emerging changes. The phenomenon of the evolving nature of data streams is known as concept drift. While there is a plethora of methods designed for detecting its occurrence, all of them assume that the drift is connected with underlying changes in the source of data. However, one must consider the possibility of a malicious injection of false data that simulates a concept drift. This adversarial setting assumes a poisoning attack that may be conducted in order to damage the underlying classification system by forcing adaptation to false data. Existing drift detectors are not capable of differentiating between real and adversarial concept drift. In this paper, we propose a framework for robust concept drift detection in the presence of adversarial and poisoning attacks. We introduce the taxonomy for two types of adversarial concept drifts, as well as a robust trainable drift detector. It is based on the augmented Restricted Boltzmann Machine with improved gradient computation and energy function. We also introduce Relative Loss of Robustness - a novel measure for evaluating the performance of concept drift detectors under poisoning attacks. Extensive computational experiments, conducted on both fully and sparsely labeled data streams, prove the high robustness and efficacy of the proposed drift detection framework in adversarial scenarios.

Recommendations

• Concept drift detection and adaptation with hierarchical hypothesis testing
• Data stream ensemble classification based on concept drift detection
• Tracking concept drift using a constrained penalized regression combiner
• An efficient and sensitive decision tree approach to mining concept-drifting data streams
• Online reliable semi-supervised learning on evolving data streams

Cites work

• \(M\)-estimation, convexity and quantiles
• Advances in Artificial Intelligence – SBIA 2004
• An efficient learning procedure for deep Boltzmann machines
• Detecting concept change in dynamic data streams
• Discovery science. 10th international conference, DS 2007, Sendai, Japan, October 1--4, 2007. Proceedings
• Estimating mutual information for feature selection in the presence of label noise
• Kappa updated ensemble for drifting data stream mining

Cited in

(1)

• Data poisoning attacks in contextual bandits

This page was built for publication: Adversarial concept drift detection under poisoning attacks for robust data stream mining

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6053814)