paper-with-me

홈 › Papers

AEGIS: Adversarial Target-Guided Retention-Data-Free Robust Concept Erasure from Diffusion Models

2026-02-06 · Fengpeng Li, Kemou Li, Qizhou Wang, Bo Han, Jiantao Zhou arxiv

Concept erasure helps stop diffusion models (DMs) from generating harmful content; but current methods face robustness retention trade off. Robustness means the model fine-tuned by concept erasure methods resists reactivation of erased concepts, even under semantically related prompts. Retention means unrelated concepts are preserved so the model's overall utility stays intact. Both are critical for concept erasure in practice, yet addressing them simultaneously is challenging, as existing works typically improve one factor while sacrificing the other. Prior work typically strengthens one while degrading the other, e.g., mapping a single erased prompt to a fixed safe target leaves class level remnants exploitable by prompt attacks, whereas retention-oriented schemes underperform against adaptive adversaries. This paper introduces Adversarial Erasure with Gradient Informed Synergy (AEGIS), a retention-data-free framework that advances both robustness and retention.

📄 PDF Abstract BibTeX arXiv:2602.06771

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

AEGIS: Adversarial Entropy-Guided Immune System -- Thermodynamic State Space Models for Zero-Day Network Evasion Detection

2026-04-02 · Vickson Ferrel arxiv

As TLS 1.3 encryption limits traditional Deep Packet Inspection (DPI), the security community has pivoted to Euclidean Transformer-based classifiers (e.g., ET-BERT) for encrypted traffic analysis. However, these models r…

AegisRF: Adversarial Perturbations Guided with Sensitivity for Protecting Intellectual Property of Neural Radiance Fields

2025-10-22 · Woo Jae Kim, Kyu Beom Han, Yoonki Cho, Youngju Na 외 arxiv

As Neural Radiance Fields (NeRFs) have emerged as a powerful tool for 3D scene representation and novel view synthesis, protecting their intellectual property (IP) from unauthorized use is becoming increasingly crucial. …

Novel View SynthesisImage Classification

AEGIS: Preserving privacy of 3D Facial Avatars with Adversarial Perturbations

2025-11-21 · Dawid Wolkiewicz, Anastasiya Pechko, Przemysław Spurek, Piotr Syga arxiv

The growing adoption of photorealistic 3D facial avatars, particularly those utilizing efficient 3D Gaussian Splatting representations, introduces new risks of online identity theft, especially in systems that rely on bi…

Face Verification

An Empirical Study of Aegis

2024-04-24 · Daniel Saragih, Paridhi Goel, Tejas Balaji, Alyssa Li

Bit flipping attacks are one class of attacks on neural networks with numerous defense mechanisms invented to mitigate its potency. Due to the importance of ensuring the robustness of these defense mechanisms, we perform…

Data Augmentation

Towards Backdoor Attacks and Defense in Robust Machine Learning Models

2020-02-25 · Ezekiel Soremekun, Sakshi Udeshi, Sudipta Chattopadhyay

The introduction of robust optimisation has pushed the state-of-the-art in defending against adversarial attacks. Notably, the state-of-the-art projected gradient descent (PGD)-based training method has been shown to be …

BIG-bench Machine LearningClustering