analiza aiqbench
This survey establishes a unified taxonomy for transferable adversarial attacks across six axes, clarifying mechanisms like gradient alignment and shared representations that allow vulnerabilities to propagate between different AI systems. By distinguishing source-target configurations from attacker budgets, the work addresses critical gaps in evaluation metrics and provides a framework for assessing security risks in model stealing, backdoors, and evasion scenarios. It sets a clear research agenda for reproducible benchmarking and transfer-aware defense strategies essential for robust AI deployment.
Beyond Boundaries: A Comprehensive Survey of Transferable Attacks on AI Systems
arXiv:2311.11796v3 Announce Type: replace-cross Abstract: Transferable attacks allow adversarial manipulations or learned attack mechanisms constructed in one setting to compromise other AI systems, potentially reducing attack costs and amplifying security and privacy risks. Yet fragmented terminology and evaluation practices hinder understanding of what transfers, why it transfers, and where transfer fails. This survey synthesizes representative research and develops a unified taxonomy spanning six complementary axes: cross-instance, cross-domain, cross-modality, cross-model, cross-task, and cross-deployment. By separating source-target configuration changes from attacker access and adaptation budgets, the taxonomy provides a common framework for comparing transfer across evasion, poisoning, backdoors, privacy inference, model stealing, and prompt-based attacks. We connect transfer mechanisms, including gradient alignment, shared representations, training persistence, and deployment invariances, to enhancement methods based on augmentation, disentanglement, and optimization. Our synthesis identifies inconsistent transfer, incompatible metrics, and gaps between simulated and physical evaluations, and establishes a research agenda for compositional transfer, configuration differences, temporal persistence, and reproducible benchmarking. Together, this work supports systematic investigation of shared vulnerabilities and guides transfer-aware security assessment and defense.
Abstract:Transferable attacks allow adversarial manipulations or learned attack mechanisms constructed in one setting to compromise other AI systems, potentially reducing attack costs and amplifying security and privacy risks. Yet fragmented terminology and evaluation practices hinder understanding of what transfers, why it transfers, and where transfer fails. This survey synthesizes representative research and develops a unified taxonomy spanning six complementary axes: cross-instance, cross-domain, cross-modality, cross-model, cross-task, and cross-deployment. By separating source-target configuration changes from attacker access and adaptation budgets, the taxonomy provides a common framework for comparing transfer across evasion, poisoning, backdoors, privacy inference, model stealing, and prompt-based attacks. We connect transfer mechanisms, including gradient alignment, shared representations, training persistence, and deployment invariances, to enhancement methods based on augmentation, disentanglement, and optimization. Our synthesis identifies inconsistent transfer, incompatible metrics, and gaps between simulated and physical evaluations, and establishes a research agenda for compositional transfer, configuration differences, temporal persistence, and reproducible benchmarking. Together, this work supports systematic investigation of shared vulnerabilities and guides transfer-aware security assessment and defense.
Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
arXiv:2311.11796 [cs.CR]
(or arXiv:2311.11796v3 [cs.CR] for this version)
https://doi.org/10.48550/arXiv.2311.11796 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 272 cuvinte
Subiecte conexe
Digital States 2026: AI Moves From Experiment to Infrastructure
govtech.com · 24 sept. 2026
News
Quantum Computing Report · 7 iun. 2026
From Acoustic Sensing to Physical Actions: A Comprehensive Survey of Security Risks in Edge-Deployed Audio LLM-Based Agents
arXiv · 14 oct. 2026
Game Design Creativity evals
LessWrong · 12 oct. 2026