Data poisoning attacks in contextual bandits
WebMay 16, 2024 · Stochastic multi-armed bandits form a class of online learning problems that have important applications in online recommendation systems, adaptive medical treatment, and many others. Even though potential attacks against these learning algorithms may hijack their behavior, causing catastrophic loss in real-world applications, little is known ... WebMay 16, 2024 · Stochastic multi-armed bandits form a class of online learning problems that have important applications in online recommendation systems, adaptive medical …
Data poisoning attacks in contextual bandits
Did you know?
WebFeb 10, 2024 · In this paper, we study several attack scenarios and show that a malicious agent can force a linear contextual bandit algorithm to pull any desired arm T - o (T) times over a horizon of T steps, while applying adversarial modifications to either rewards or contexts that only grow logarithmically as O (log T). We also investigate the case when a ... WebWe study offline data poisoning attacks in contextual bandits, a class of reinforcement learning problems with important applications in online recommendation and adaptive medical treatment, among others. We provide a general attack framework …
WebDec 10, 2024 · In order to develop trustworthy contextual bandit systems, understanding the impacts of various adversarial attacks on contextual bandit algorithms is essential. … WebDec 1, 2024 · By using a novel contextual multi-armed bandit-based algorithm to balance the accuracy and efficiency, we propose a context-based online big-data-oriented …
WebFeb 10, 2024 · In this paper, we study several attack scenarios and show that a malicious agent can force a linear contextual bandit algorithm to pull any desired arm T - o (T) … WebUpload an image to customize your repository’s social media preview. Images should be at least 640×320px (1280×640px for best display).
WebMar 30, 2024 · 攻击方法：. 1）Functional Adversarial Attacks 2）Improving Black-box Adversarial Attacks with a Transfer-based Prior 3）Cross-Domain Transferability of Adversarial Perturbations 4）Subspace Attack: Exploiting Promising Subspaces for Query-Efficient Black-box Attacks 5）A Unified Framework for Data Poisoning Attack to …
WebAug 17, 2024 · We study offline data poisoning attacks in contextual bandits, a class of reinforcement learning problems with important applications in online recommendation … can chickens eat cherries with pitsfish in soap dispenserWebData Poisoning Attacks in Contextual Bandits 3 Formally, a contextual bandit has a set Xof contexts and a set A= f1;2;:::;Kgof K arms. A contextual bandit algorithm proceeds … fish in snellvilleWebon when and where the attack happens. In a typical data poisoning attack (a.k.a. training-time attack) setting, the attacker tampers the training data during training time to downgrade the utility of the learned model. On the other hand, in adver-sarial examples (a.k.a test-time attack), the attacker manipulates features of a target can chickens eat cherry pitsWebSep 26, 2024 · Data Poisoning Attacks in Contextual Bandits: 9th International Conference, GameSec 2024, Seattle, WA, USA, October 29–31, 2024, Proceedings September 2024 DOI: 10.1007/978-3-030-01554-1_11 can chickens eat chipsWebIn this paper, we study the action poisoning attack against linear contextual bandit in both white-box and black-box settings. In the white-box setting, we assume that the attacker knows the coefficient vectors associated with arms. Thus, at each round, the attacker knows the mean rewards of all arms. While it is often unrealistic to exactly know fish in soupWebFeb 10, 2024 · Adversarial Attacks on Linear Contextual Bandits. Contextual bandit algorithms are applied in a wide range of domains, from advertising to recommender … can chickens eat chicory