paper-with-me

홈 › Papers

A Kernelized Manifold Mapping to Diminish the Effect of Adversarial Perturbations

2019-03-03 · CVPR 2019 6 · Saeid Asgari Taghanaki, Kumar Abhishek, Shekoofeh Azizi, Ghassan Hamarneh

The linear and non-flexible nature of deep convolutional models makes them vulnerable to carefully crafted adversarial perturbations. To tackle this problem, we propose a non-linear radial basis convolutional feature mapping by learning a Mahalanobis-like distance function. Our method then maps the convolutional features onto a linearly well-separated manifold, which prevents small adversarial perturbations from forcing a sample to cross the decision boundary. We test the proposed method on three publicly available image classification and segmentation datasets namely, MNIST, ISBI ISIC 2017 skin lesion segmentation, and NIH Chest X-Ray-14. We evaluate the robustness of our method to different gradient (targeted and untargeted) and non-gradient based attacks and compare it to several non-gradient masking defense strategies. Our results demonstrate that the proposed method can increase the resilience of deep convolutional neural networks to adversarial perturbations without accuracy drop on clean data.

📄 PDF Abstract BibTeX arXiv:1903.01015

Code (1)

asgsaeid/KernelizedManifoldMapping 공식 구현

Tasks

image-classificationImage ClassificationLesion SegmentationSegmentationSkin Lesion Segmentation

Similar Papers 제목 키워드 기반

Action Anticipation with RBF Kernelized Feature Mapping RNN

2019-11-18 · ECCV 2018 9 · Yuge Shi, Basura Fernando, Richard Hartley

We introduce a novel Recurrent Neural Network-based algorithm for future video feature generation and action anticipation called feature mapping RNN. Our novel RNN architecture builds upon three effective principles of m…

Action Anticipation

Extrinsic Methods for Coding and Dictionary Learning on Grassmann Manifolds

2014-01-31 · Mehrtash Harandi, Richard Hartley, Chunhua Shen, Brian Lovell 외

Sparsity-based representations have recently led to notable results in various visual recognition tasks. In a separate line of research, Riemannian manifolds have been shown useful for dealing with features and models th…

Action RecognitionClassificationDictionary LearningFace Recognition+4

Kernelized LRR on Grassmann Manifolds for Subspace Clustering

2016-01-09 · Boyue Wang, Yongli Hu, Junbin Gao, Yanfeng Sun 외

Low rank representation (LRR) has recently attracted great interest due to its pleasing efficacy in exploring low-dimensional sub- space structures embedded in data. One of its successful applications is subspace cluster…

Clustering

Kernelized Low Rank Representation on Grassmann Manifolds

2015-04-08 · Boyue Wang, Yongli Hu, Junbin Gao, Yanfeng Sun 외

Low rank representation (LRR) has recently attracted great interest due to its pleasing efficacy in exploring low-dimensional subspace structures embedded in data. One of its successful applications is subspace clusterin…

Clustering

Manifold Regularization for Kernelized LSTD

2017-10-15 · Xinyan Yan, Krzysztof Choromanski, Byron Boots, Vikas Sindhwani

Policy evaluation or value function or Q-function approximation is a key procedure in reinforcement learning (RL). It is a necessary component of policy iteration and can be used for variance reduction in policy gradient…

Policy Gradient MethodsReinforcement LearningReinforcement Learning (RL)