Filters








16,283 Hits in 5.9 sec

Searching towards Class-Aware Generators for Conditional Generative Adversarial Networks [article]

Peng Zhou, Lingxi Xie, Xiaopeng Zhang, Bingbing Ni, Qi Tian
2021 arXiv   pre-print
Conditional Generative Adversarial Networks (cGAN) were designed to generate images based on the provided conditions, \eg, class-level distributions.  ...  The search space contains regular and class-modulated convolutions, where the latter is designed to introduce class-specific information while avoiding the reduction of training data for each class generator  ...  That being said, the class-aware generators should be searched under the condition that the discriminator is also class-aware.  ... 
arXiv:2006.14208v2 fatcat:qrygtqfo6vgxpiowuyjwza27ye

Sandwich Batch Normalization: A Drop-In Replacement for Feature Distribution Heterogeneity [article]

Xinyu Gong, Wuyang Chen, Tianlong Chen, Zhangyang Wang
2021 arXiv   pre-print
We demonstrate the prevailing effectiveness of SaBN as a drop-in replacement in four tasks: conditional image generation, neural architecture search (NAS), adversarial training, and arbitrary style transfer  ...  weight-sharing NAS algorithm significantly on NAS-Bench-201; substantially improves the robust and standard accuracies for adversarial defense; and produces superior arbitrary stylized results.  ...  networks; Group Normalization (GN) [60] for tackling small batch sizes; StochNorm [36] for fine-tuning; Passport-aware Normalization [72] for model IP protection; and [38, 58, 73] for image generation  ... 
arXiv:2102.11382v2 fatcat:hf4rr73j4fdfxhpyywvhfzcena

Deflecting Adversarial Attacks [article]

Yao Qin, Nicholas Frosst, Colin Raffel, Garrison Cottrell, Geoffrey Hinton
2020 arXiv   pre-print
We present a new approach towards ending this cycle where we "deflect" adversarial attacks by causing the attacker to produce an input that semantically resembles the attack's target class.  ...  These attack images can no longer be called "adversarial" because our network classifies them the same way as humans do.  ...  An example of a clean input, an adversarial example generated via a PGD attack, and the reconstructions for the clean and adversarial inputs from each class capsule.  ... 
arXiv:2002.07405v1 fatcat:o4e3p3qbqzfj3ecviv7ggml5ta

On the Need for Topology-Aware Generative Models for Manifold-Based Defenses [article]

Uyeong Jang, Susmit Jha, Somesh Jha
2020 arXiv   pre-print
In this paper, we investigate the following question: do the generative models used in manifold-based defenses need to be topology-aware?  ...  The existence of adversarial examples has hindered the deployment of ML algorithms in safety-critical sectors, such as security. Several defenses for adversarial examples exist in the literature.  ...  For any dataset, the class-aware INC achieves lower projection errors.  ... 
arXiv:1909.03334v4 fatcat:dc6vzg2u6benjkszw6nn6qm5ti

Detecting and Diagnosing Adversarial Images with Class-Conditional Capsule Reconstructions [article]

Yao Qin, Nicholas Frosst, Sara Sabour, Colin Raffel, Garrison Cottrell, Geoffrey Hinton
2020 arXiv   pre-print
In this paper, we first detect adversarial examples or otherwise corrupted images based on a class-conditional reconstruction of the input.  ...  Then, we diagnose the adversarial examples for CapsNets and find that the success of the reconstructive attack is highly related to the visual similarity between the source and target class.  ...  Instead, in this paper we develop methods for detecting adversarial examples by making use of class-conditional reconstruction networks.  ... 
arXiv:1907.02957v2 fatcat:xcxwzrth5jhmnbtrxcxvfuy2gu

A survey on Adversarial Recommender Systems: from Attack/Defense strategies to Generative Adversarial Networks [article]

Yashar Deldjoo and Tommaso Di Noia and Felice Antonio Merra
2020 arXiv   pre-print
successful application of AML in generative adversarial networks (GANs) for generative applications, thanks to their ability for learning (high-dimensional) data distributions.  ...  This review serves as a reference for the RS community, working on the security of RS or on generative models using GANs to improve their quality.  ...  on class conditioning on both the generator and discriminator) [100] .  ... 
arXiv:2005.10322v2 fatcat:4wqcluqgnbbwpkicunn42et5te

SentiNet: Detecting Localized Universal Attacks Against Deep Learning Systems [article]

Edward Chou, Florian Tramèr, Giancarlo Pellegrino
2020 arXiv   pre-print
SentiNet is a novel detection framework for localized universal attacks on neural networks.  ...  These attacks restrict adversarial noise to contiguous portions of an image and are reusable with different images -- constraints that prove useful for generating physically-realizable attacks.  ...  Finally, SentiNet uses Tensorflow 1.5 to generate the adversarial patches for the uncompromised network, BLVC-Caffe for the trojaned network, and Faster-RCNN Caffe [17] for the poisoned network.  ... 
arXiv:1812.00292v4 fatcat:5vg4nuit2vdenmkzlrinotqppe

A survey on generative adversarial networks for imbalance problems in computer vision tasks

Vignesh Sampath, Iñaki Maurtua, Juan José Aguilar Martín, Aitor Gutierrez
2021 Journal of Big Data  
In recent years, Generative Adversarial Neural Networks (GANs) have gained immense attention by researchers across a variety of application domains due to their capability to model complex real-world image  ...  It is particularly important that GANs can not only be used to generate synthetic images, but also its fascinating adversarial learning idea showed good potential in restoring balance in imbalanced datasets  ...  Acknowledgements The authors would like to thank the anonymous reviewers for their valuable comments and suggestions on the paper.  ... 
doi:10.1186/s40537-021-00414-0 pmid:33552840 pmcid:PMC7845583 fatcat:g3p6hbjuj5c5vbe23ms4g6ed6q

Boundless: Generative Adversarial Networks for Image Extension [article]

Piotr Teterwak, Aaron Sarna, Dilip Krishnan, Aaron Maschinot, David Belanger, Ce Liu, William T. Freeman
2019 arXiv   pre-print
We introduce semantic conditioning to the discriminator of a generative adversarial network (GAN), and achieve strong results on image extension with coherent semantics and visually pleasing colors and  ...  We also show promising results in extreme extensions, such as panorama generation.  ...  Huiwen Chang for helpful discussion and sharing code, and Dr. Guilin Liu for helping with running Partial Convolution comparison experiments.  ... 
arXiv:1908.07007v1 fatcat:fv3a2j6bprcffm5cibtsoicxam

Towards the first adversarially robust neural network model on MNIST [article]

Lukas Schott, Jonas Rauber, Matthias Bethge, Wieland Brendel
2018 arXiv   pre-print
We present a novel robust classification model that performs analysis by synthesis using learned class-conditional data distributions.  ...  the perceptual boundary between the original and the adversarial class.  ...  Finally, adversarials generated for the ABS models are semantically meaningful for humans and are sitting close to the perceptual boundary between the original and the adversarial class.  ... 
arXiv:1805.09190v3 fatcat:f5wvcvsomncvnabtyspg6lsmwu

Editorial: Special Issue on Deep Learning for Face Analysis

Chen Change Loy, Xiaoming Liu, Tae-Kyun Kim, Fernando De la Torre, Rama Chellappa
2019 International Journal of Computer Vision  
Most of the aforementioned methods employ Generative Adversarial Network (GAN), which is also used by the following work for face synthesis and image-to-image translation.  ...  The generator is conditioned on a statistical shape prior with differentiable canonical shape normalisation. This enables the generation of images with realistic texture and shape.  ... 
doi:10.1007/s11263-019-01179-z fatcat:tp7uyd22bbfatg7f7wq4pafaze

Adversarial Examples in Modern Machine Learning: A Review [article]

Rey Reza Wiyatno, Anqi Xu, Ousmane Dia, Archy de Berker
2019 arXiv   pre-print
In this survey, we focus on machine learning models in the visual domain, where methods for generating and detecting such examples have been most extensively studied.  ...  We explore a variety of adversarial attack methods that apply to image-space content, real world adversarial attacks, adversarial defenses, and the transferability property of adversarial examples.  ...  The training of the detector network involves generating adversarial examples to be part of the training set for the detector network.  ... 
arXiv:1911.05268v2 fatcat:majzak4sqbhcpeahghh6sm3dwq

Task Specific Visual Saliency Prediction with Memory Augmented Conditional Generative Adversarial Networks [article]

Tharindu Fernando, Simon Denman, Sridha Sridharan, Clinton Fookes
2018 arXiv   pre-print
To address this limitation, we propose a novel saliency estimation model which leverages the semantic modelling power of conditional generative adversarial networks together with memory architectures which  ...  Our studies not only shed light on a novel application area for generative adversarial networks, but also emphasise the importance of task specific saliency modelling and demonstrate the plausibility of  ...  Generative Adversarial Networks Generative adversarial networks (GAN), which belong to the family of generative models, have achieved promising results for pixel-to-pixel synthesis [41] .  ... 
arXiv:1803.03354v1 fatcat:xtcc4agytndphaqcj3mbqvfmdu

Channel-Aware Adversarial Attacks Against Deep Learning-Based Wireless Signal Classifiers [article]

Brian Kim, Yalin E. Sagduyu, Kemal Davaslioglu, Tugba Erpek, Sennur Ulukus
2021 arXiv   pre-print
A deep neural network is used at each receiver to classify its over-the-air received signals to modulation types.  ...  The major vulnerability of modulation classifiers to over-the-air adversarial attacks is shown by accounting for different levels of information available about the channel, the transmitter input, and  ...  Channel Inversion Attack non-targeted FGM attack, the adversary searches for an attack that causes any misclassification (independent of target class).  ... 
arXiv:2005.05321v3 fatcat:bl5bgamxcrbrzm2thyr6fp56fu

Context-Aware Transfer Attacks for Object Detection [article]

Zikui Cai, Xinxin Xie, Shasha Li, Mingjun Yin, Chengyu Song, Srikanth V. Krishnamurthy, Amit K. Roy-Chowdhury, M. Salman Asif
2021 arXiv   pre-print
In this paper, we present a new approach to generate context-aware attacks for object detectors.  ...  This makes such detectors inherently context-aware and adversarial attacks in this space are more challenging than those targeting image classifiers.  ...  Similarly, size graph models the 2D distribu- Next, we show how we compose context-aware attack tions of object height and width, where edge ei,j represents plans and search for adversarial  ... 
arXiv:2112.03223v1 fatcat:nz5u7akuffebjdztn73jc37b2y
« Previous Showing results 1 — 15 out of 16,283 results