Research map / Adversarial robustness and security / Black-box and transfer attacks

Dynamic Neural Fortresses: An Adaptive Shield for Model Extraction Defense

ICLR 2025 · Adversarial robustness and security › Black-box and transfer attacks · approach: Data-free substitute training

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a defense against model extraction attacks that also protects the API's neural network architecture as intellectual property; only the abstract was available, so details are limited.

Problem

Model extraction attacks steal functionality of black-box pre-trained models behind APIs; existing defenses focus on functionality, while the API's neural network architecture as intellectual property also needs protection.

Approach

Abstract is truncated before the proposed method is described; the stated gap is that current defenses ignore architecture protection, which this work targets.

Closest papers by meaning

• Train to Defend: First Defense Against Cryptanalytic Neural Network Parameter Extraction Attacks (NeurIPS 2025 · similarity 0.88)
• HoneypotNet: Backdoor Attacks Against Model Extraction (AAAI 2025 · similarity 0.88)
• Beyond Slow Signs in High-fidelity Model Extraction (NeurIPS 2024 · similarity 0.85)
• Efficient Model Stealing Defense with Noise Transition Matrix (CVPR 2024 · similarity 0.85)
• CEGA: A Cost-Effective Approach for Graph-Based Model Extraction and Acquisition (ICML 2025 · similarity 0.85)
• SAME: Sample Reconstruction against Model Extraction Attacks (AAAI 2024 · similarity 0.84)