Image de couverture de livre

Beyond Training-time Poisoning: Component-level and Post-training Backdoors in Deep Reinforcement Learning

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Vyas, Sanyam, Caron, Alberto, Hicks, Chris, Burnap, Pete, Mavroudis, Vasilios
Format:	Preprint
Publié:	2025
Sujets:	Machine Learning Artificial Intelligence Cryptography and Security
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Mitigating Deep Reinforcement Learning Backdoors in the Neural Activation Space
par: Vyas, Sanyam, et autres
Publié: (2024)

Entity-based Reinforcement Learning for Autonomous Cyber Defence
par: Thompson, Isaac Symes, et autres
Publié: (2024)

SoK: The Pitfalls of Deep Reinforcement Learning for Cybersecurity
par: McFadden, Shae, et autres
Publié: (2026)

Autonomous Network Defence using Reinforcement Learning
par: Foley, Myles, et autres
Publié: (2024)

Less is more? Rewards in RL for Cyber Defence
par: Bates, Elizabeth, et autres
Publié: (2025)

CybORG++: An Enhanced Gym for the Development of Autonomous Cyber Agents
par: Emerson, Harry, et autres
Publié: (2024)

HonestCyberEval: An AI Cyber Risk Benchmark for Automated Software Exploitation
par: Ristea, Dan, et autres
Publié: (2024)

Referential Security as a New Paradigm for AI Evaluations
par: Ristea, Dan, et autres
Publié: (2026)

DRMD: Deep Reinforcement Learning for Malware Detection under Concept Drift
par: McFadden, Shae, et autres
Publié: (2025)

An Attentive Graph Agent for Topology-Adaptive Cyber Defence
par: Sandoval, Ilya Orson, et autres
Publié: (2025)

Zero-Trust Network Access (ZTNA)
par: Mavroudis, Vasilios
Publié: (2024)

What if we could hot swap our Biometrics?
par: Crowcroft, Jon, et autres
Publié: (2025)

Analysis of Publicly Accessible Operational Technology and Associated Risks
par: Rodda, Matthew, et autres
Publié: (2025)

Quantifying Mix Network Privacy Erosion with Generative Models
par: Mavroudis, Vasilios, et autres
Publié: (2025)

From Promise to Peril: Rethinking Cybersecurity Red and Blue Teaming in the Age of LLMs
par: Abuadbba, Alsharif, et autres
Publié: (2025)

One Pic is All it Takes: Poisoning Visual Document Retrieval Augmented Generation with a Single Image
par: Shereen, Ezzeldin, et autres
Publié: (2025)

Consistent and Compatible Modelling of Cyber Intrusions and Incident Response Demonstrated in the Context of Malware Attacks on Critical Infrastructure
par: Maynard, Peter, et autres
Publié: (2025)

Beyond Rewards in Reinforcement Learning for Cyber Defence
par: Bates, Elizabeth, et autres
Publié: (2026)

A Set of Generalized Components to Achieve Effective Poison-only Clean-label Backdoor Attacks with Collaborative Sample Selection and Triggers
par: Wu, Zhixiao, et autres
Publié: (2025)

Backdoor Learning Curves: Explaining Backdoor Poisoning Beyond Influence Functions
par: Cinà, Antonio Emanuele, et autres
Publié: (2021)

From Poisoned to Aware: Fostering Backdoor Self-Awareness in LLMs
par: Shen, Guangyu, et autres
Publié: (2025)

CBPF: Filtering Poisoned Data Based on Composite Backdoor Attack
par: Xia, Hanfeng, et autres
Publié: (2024)

PBP: Post-training Backdoor Purification for Malware Classifiers
par: Nguyen, Dung Thuy, et autres
Publié: (2024)

On Efficient Bayesian Exploration in Model-Based Reinforcement Learning
par: Caron, Alberto, et autres
Publié: (2025)

Data Poisoning in Deep Learning: A Survey
par: Zhao, Pinlong, et autres
Publié: (2025)

Empowering IoT Security: On-Device Intrusion Detection in Resource Constrained Devices
par: Ieropoulos, Vasilis, et autres
Publié: (2026)

BadSkill: Backdoor Attacks on Agent Skills via Model-in-Skill Poisoning
par: Tie, Guiyao, et autres
Publié: (2026)

Toward Polymorphic Backdoor against Semantic Communication via Intensity-Based Poisoning
par: Yang, Xiao, et autres
Publié: (2026)

A Spatiotemporal Stealthy Backdoor Attack against Cooperative Multi-Agent Deep Reinforcement Learning
par: Yu, Yinbo, et autres
Publié: (2024)

P2P: A Poison-to-Poison Remedy for Reliable Backdoor Defense in LLMs
par: Zhao, Shuai, et autres
Publié: (2025)

Concealing Backdoor Model Updates in Federated Learning by Trigger-Optimized Data Poisoning
par: Zhang, Yujie, et autres
Publié: (2024)

Persistent Pre-Training Poisoning of LLMs
par: Zhang, Yiming, et autres
Publié: (2024)

Poison Once, Control Anywhere: Clean-Text Visual Backdoors in VLM-based Mobile Agents
par: Wang, Xuan, et autres
Publié: (2025)

Class-Conditional Neural Polarizer: A Lightweight and Effective Backdoor Defense by Purifying Poisoned Features
par: Zhu, Mingli, et autres
Publié: (2025)

The Stronger the Diffusion Model, the Easier the Backdoor: Data Poisoning to Induce Copyright Breaches Without Adjusting Finetuning Pipeline
par: Wang, Haonan, et autres
Publié: (2024)

Angel or Demon: Investigating the Plasticity Interventions' Impact on Backdoor Threats in Deep Reinforcement Learning
par: Ma, Oubo, et autres
Publié: (2026)

UNIDOOR: A Universal Framework for Action-Level Backdoor Attacks in Deep Reinforcement Learning
par: Ma, Oubo, et autres
Publié: (2025)

SleeperNets: Universal Backdoor Poisoning Attacks Against Reinforcement Learning Agents
par: Rathbun, Ethan, et autres
Publié: (2024)

BackdoorMBTI: A Backdoor Learning Multimodal Benchmark Tool Kit for Backdoor Defense Evaluation
par: Yu, Haiyang, et autres
Publié: (2024)

DeBackdoor: A Deductive Framework for Detecting Backdoor Attacks on Deep Models with Limited Data
par: Popovic, Dorde, et autres
Publié: (2025)