Výsledky vyhledávání - "Zhu, Mingkang"

Report

LogoSticker: Inserting Logos into Diffusion Models for Customized Generation

Autor: Zhu, Mingkang, Chen, Xi, Wang, Zhongdao, Zhao, Hengshuang, Jia, Jiaya

Recent advances in text-to-image model customization have underscored the importance of integrating new concepts with a few examples. Yet, these progresses are largely confined to widely recognized subjects, which can be learned with relative ease th

Externí odkaz: http://arxiv.org/abs/2407.13752

Zobrazit plný text záznamu

Report

Learning to Generalize Provably in Learning to Optimize

Autor: Yang, Junjie, Chen, Tianlong, Zhu, Mingkang, He, Fengxiang, Tao, Dacheng, Liang, Yingbin, Wang, Zhangyang

Learning to optimize (L2O) has gained increasing popularity, which automates the design of optimizers by data-driven approaches. However, current L2O methods often suffer from poor generalization performance in at least two folds: (i) applying the L2

Externí odkaz: http://arxiv.org/abs/2302.11085

Zobrazit plný text záznamu

Report

Sparse and Imperceptible Adversarial Attack via a Homotopy Algorithm

Autor: Zhu, Mingkang, Chen, Tianlong, Wang, Zhangyang

Sparse adversarial attacks can fool deep neural networks (DNNs) by only perturbing a few pixels (regularized by l_0 norm). Recent efforts combine it with another l_infty imperceptible on the perturbation magnitudes. The resultant sparse and impercept

Externí odkaz: http://arxiv.org/abs/2106.06027

Zobrazit plný text záznamu

Report

Learning Transferable 3D Adversarial Cloaks for Deep Trained Detectors

Autor: Maesumi, Arman, Zhu, Mingkang, Wang, Yi, Chen, Tianlong, Wang, Zhangyang, Bajaj, Chandrajit

This paper presents a novel patch-based adversarial attack pipeline that trains adversarial patches on 3D human meshes. We sample triangular faces on a reference human mesh, and create an adversarial texture atlas over those faces. The adversarial te

Externí odkaz: http://arxiv.org/abs/2104.11101

Zobrazit plný text záznamu