InfoResearchPeer-reviewed
TRUE: Towards Robust, Unique, and Efficient Model Fingerprint Generation and Verification
- Published
- Record updated
Summary
TRUE is a two-stage framework for authenticating deep neural network models against model stealing. It generates adversarial examples as fingerprints, hardened with a minmax robust optimization strategy, then weights them with a Boltzmann distribution and samples by those weights to cut query counts. The authors report superiority across real-world datasets and several model architectures.
Topics
Related items
- InfoAntidistillation is Cheap: Lightweight Distillation-Resistance via Stylistic CorruptionSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
- InfoCutpurse: Decomposing Explanation-Guided Model Extraction on Tabular DataSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
- InfoFingerprinting the Dispatchers behind Distributed Model ExtractionSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
- InfoModel Extraction of Convolutional Neural Networks with Max-PoolingSimilar attack · OpenAlex (peer-reviewed AI security)
- InfoTrustworthy Distributed Machine Learning in Cloud-Edge-IoT Systems: A Heterogeneity-Aware Survey of Security, Privacy, and RobustnessSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)