SOTAVerified

Model extraction

Model extraction attacks, aka model stealing attacks, are used to extract the parameters from the target model. Ideally, the adversary will be able to steal and replicate a model that will have a very similar performance to the target model.

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 51–100 of 176 papers

TitleStatusHype
Marich: A Query-efficient Distributionally Equivalent Model Extraction Attack using Public DataCode0
Process Extraction from Text: Benchmarking the State of the Art and Paving the Way for Future ChallengesCode0
Protecting Intellectual Property of Language Generation APIs with Lexical WatermarkCode0
MeaeQ: Mount Model Extraction Attacks with Efficient QueriesCode0
FLuID: Mitigating Stragglers in Federated Learning using Invariant DropoutCode0
Safe and Robust Watermark Injection with a Single OoD ImageCode0
SAME: Sample Reconstruction against Model Extraction AttacksCode0
Fraternal Twins: Unifying Attacks on Machine Learning and Digital Watermarking—0
GENIE: Watermarking Graph Neural Networks for Link Prediction—0
Good Artists Copy, Great Artists Steal: Model Extraction Attacks Against Image Translation Models—0
Grey-box Extraction of Natural Language Models—0
GrOVe: Ownership Verification of Graph Neural Networks using Embeddings—0
HODA: Hardness-Oriented Detection of Model Extraction Attacks—0
High Accuracy and High Fidelity Extraction of Neural Networks—0
HODA: Protecting DNNs Against Model Extraction Attacks via Hardness of Samples—0
HoneypotNet: Backdoor Attacks Against Model Extraction—0
Increasing the Cost of Model Extraction with Calibrated Proof of Work—0
Interpretability via Model Extraction—0
Interpreting Blackbox Models via Model Extraction—0
Killing One Bird with Two Stones: Model Extraction and Attribute Inference Attacks against BERT-based APIs—0
Noisy Data Meets Privacy: Training Local Models with Post-Processed Remote Queries—0
Learnable Linguistic Watermarks for Tracing Model Extraction Attacks on Large Language Models—0
Leveraging Extracted Model Adversaries for Improved Black Box Attacks—0
Like an Open Book? Read Neural Network Architecture with Simple Power Analysis on 32-bit Microcontrollers—0
MaskedNet: The First Hardware Inference Engine Aiming Power Side-Channel Protection—0
MEAOD: Model Extraction Attack against Object Detectors—0
MEGEX: Data-Free Model Extraction Attack against Gradient-Based Explainable AI—0
Mercury: An Automated Remote Side-channel Attack to Nvidia Deep Learning Accelerator—0
Mitigating Query-Flooding Parameter Duplication Attack on Regression Models with High-Dimensional Gaussian Mechanism—0
Model Extraction and Adversarial Attacks on Neural Networks using Switching Power Information—0
Model Extraction and Defenses on Generative Adversarial Networks—0
Model Extraction Attack against Self-supervised Speech Models—0
Model Extraction Attacks Against Reinforcement Learning Based Controllers—0
Model Extraction Attacks against Recurrent Neural Networks—0
Model Extraction Attacks on Split Federated Learning—0
Model Extraction Attacks Revisited—0
Model Extraction Warning in MLaaS Paradigm—0
Monitoring-based Differential Privacy Mechanism Against Query-Flooding Parameter Duplication Attack—0
NASPY: Automated Extraction of Automated Machine Learning Models—0
NaturalFinger: Generating Natural Fingerprint with Generative Adversarial Networks—0
Navigating the Deep: Signature Extraction on Deep Neural Networks—0
On the amplification of security and privacy risks by post-hoc explanations in machine learning models—0
On the interplay of Explainability, Privacy and Predictive Performance with Explanation-assisted Model Extraction—0
Ownership Protection of Generative Adversarial Networks—0
Pareto-Secure Machine Learning (PSML): Fingerprinting and Securing Inference Serving Systems—0
Power-Based Attacks on Spatial DNN Accelerators—0
Precise Extraction of Deep Learning Models via Side-Channel Attacks on Edge/Endpoint Devices—0
Privacy Implications of Explainable AI in Data-Driven Systems—0
ProDiF: Protecting Domain-Invariant Features to Secure Pre-Trained Models Against Extraction—0
Protecting Copyright of Medical Pre-trained Language Models: Training-Free Backdoor Model Watermarking—0
Show:102550
← PrevPage 2 of 4Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1three-step-originalExact Match0.17—Unverified