14,411 Hits in 6.4 sec

Studying Stability of Different Convolutional Neural Networks Against Additive Noise

Hamed H. Aghdam, Elnaz J. Heravi, Domenec Puig
2017 Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications  
However, these techniques do not usually provide a tool for estimating stability of a ConvNet against noise. In this paper, we show how to analyze a ConvNet in the frequency domain.  ...  the effect of high frequencies.They also show that a convolution kernel with more concentrated frequency response is more stable against noise.  ...  Studying Stability of Different Convolutional Neural Networks Against Additive Noise.  ... 
doi:10.5220/0006200003620369 dblp:conf/visapp/AghdamHP17b fatcat:5tcux27sg5gkjb3s3fgl7l3ntq

Analyzing Stability of Convolutional Neural Networks in the Frequency Domain [article]

Elnaz J. Heravi, Hamed H. Aghdam, Domenec Puig
2015 arXiv   pre-print
However, these techniques do not usually provide a tool for estimating the stability of a ConvNet against noise.  ...  Our experiments on a few ConvNets trained on different datasets revealed that convolution kernels of a trained ConvNet usually pass most of the frequencies and they are not able to effectively eliminate  ...  CONCLUSION In this paper, we studied the stability of Convolutional Neural Networks (ConvNets) against image degradation.  ... 
arXiv:1511.03042v2 fatcat:rh7dgm44l5eqbompk3tvfy5yla

Using a Deep Quantum Neural Network to Enhance the Fidelity of Quantum Convolutional Codes

Hanwei Xiao, Xiaoguang Chen, Jin Xu
2022 Applied Sciences  
Towards the circuit of quantum convolutional codes, the target quantum state |0⟩ or |1⟩ is turned into entangled quantum states, which can defend against quantum noise more effectively.  ...  In this paper, we use a quantum neural network (QNN) to enhance the fidelity of [6,2,2] quantum convolutional codes.  ...  Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/app12115662 fatcat:we3ezt6l6res5gatmpfc2zv67y

Fault Tolerance of Neural Networks in Adversarial Settings [article]

Vasisht Duddu, N. Rajesh Pillai, D. Vijay Rao, Valentina E. Balas
2019 arXiv   pre-print
Specifically, this work studies the impact of the fault tolerance of the Neural Network on training the model by adding noise to the input (Adversarial Robustness) and noise to the gradients (Differential  ...  Artificial Intelligence systems require a through assessment of different pillars of trust, namely, fairness, interpretability, data and model privacy, reliability (safety) and robustness against against  ...  In case of FashionMNIST dataset, two different neural network architectures are used, a MultiLayer Perceptron and a Convolutional Neural Network.  ... 
arXiv:1910.13875v1 fatcat:vk6fxfb2rrdapamwsehbzt2bku

Enhancing the Performance of Convolutional Neural Networks on Quality Degraded Datasets [article]

Jonghwa Yim, Kyung-Ah Sohn
2017 arXiv   pre-print
Abnormal factors, including real-world noise, blur, or other quality degradations, ruin the output of a neural network.  ...  Despite the appeal of deep neural networks that largely replace the traditional handmade filters, they still suffer from isolated cases that cannot be properly handled only by the training of convolutional  ...  For example, the studies [3] , [9] focused on the fine-tuning of the neural network model to make it robust to noise, and others, such as [4] , attempted to modify the design of the neural network  ... 
arXiv:1710.06805v1 fatcat:3v4bsrbgynfnbhd6pdsikp3ogi

Investigations of the Systematic Uncertainties in Convolutional Neural Network Based Analysis of Atmospheric Cherenkov Telescope Data [article]

R.D. Parsons, A.M.W. Mitchell, S. Ohm
2022 arXiv   pre-print
However, a strong dependence of the neural network predictions on the noise level within the camera was found, with differences of up to 50% in the gamma-ray acceptance rate in very noisy environments.  ...  We investigate the stability of convolutional recurrent networks by applying them to background rejection in a toy Monte Carlo simulation of a Cherenkov telescope array.  ...  Discussion The study of changing observing conditions on the behaviour of convolutional neural networks used for background rejections in IACTs has shown several interesting effects on the network performance  ... 
arXiv:2203.05315v1 fatcat:3idf5ruqr5g35d3cuedm2tu3gm

Identification of hydrodynamic instability by convolutional neural networks [article]

Wuyue Yang, Liangrong Peng, Yi Zhu, Liu Hong
2020 arXiv   pre-print
In this paper, modern machine learning techniques, especially the convolutional neural networks (CNN), are applied to identify the transition between different flow motions raised by hydrodynamic instability  ...  In addition, key spatial features used for classifying different flow patterns are revealed by the principal component analysis.  ...  ACKNOWLEDGEMENTS Declaration of Interests. The authors declare no conflicts of interest.  ... 
arXiv:2006.01446v1 fatcat:f5ryskeobjfylcgfeayaiwhdae

Towards calibration-invariant spectroscopy using deep learning

M. Chatzidakis, G. A. Botton
2019 Scientific Reports  
Here we propose automated feature extraction using deep convolutional neural networks to determine the class of compound given only the shape of the spectrum.  ...  We test a variety of commonly used neural network architectures found in the literature and propose a new fully convolutional architecture with improved translation-invariance which is immune to calibration  ...  In this study, we create a new method of spectroscopy analysis using deep convolutional neural networks.  ... 
doi:10.1038/s41598-019-38482-1 pmid:30765890 pmcid:PMC6376024 fatcat:n4mwduuoozelbhyofikgdkjr7q

How Does Noise Help Robustness? Explanation and Exploration under the Neural SDE Framework

Xuanqing Liu, Tesi Xiao, Si Si, Qin Cao, Sanjiv Kumar, Cho-Jui Hsieh
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
For regularization purposes, our framework includes multiple types of noise patterns, such as dropout, additive, and multiplicative noise, which are common in plain neural networks.  ...  We provide some theoretical analyses explaining the improved robustness of our models against input perturbations.  ...  In this paper, we study the role of randomness for training a robust neural network.  ... 
doi:10.1109/cvpr42600.2020.00036 dblp:conf/cvpr/LiuXSCKH20 fatcat:2kpebczdbbfp7jvfy3gafotdpi

Intelligent Fault Diagnosis Method for Blade Damage of Quad-Rotor UAV Based on Stacked Pruning Sparse Denoising Autoencoder and Convolutional Neural Network

Pu Yang, Chenwan Wen, Huilin Geng, Peng Liu
2021 Machines  
The convolutional neural network model with sPSDAE has a faster training speed, strong adaptability to noise interference signals, and can also suppress the over-fitting problem of the convolutional neural  ...  neural networks.  ...  Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/machines9120360 fatcat:f73ycymkbjbvfn6eq4uybhjmei

Comparison of discrete transforms for deep‐neural‐networks‐based speech enhancement

Wissam A. Jassim, Naomi Harte
2022 IET Signal Processing  
In addition to the traditional FNN, recurrent neural networks (RNN) and convolutional neural networks (CNN) were tested in [2] .  ...  Two deep-learning architectures were tested: convolutional neural networks (CNN) and fully connected neural networks.  ...  T A B L E 2 Detailed information about the network layers of the architecture of the convolutional neural networks topology.  ... 
doi:10.1049/sil2.12109 fatcat:y2qvnist35gw3lauvmj2lhc46a

Determination of the Semion Code Threshold using Neural Decoders [article]

Santiago Varona, Miguel Angel Martin-Delgado
2020 arXiv   pre-print
Thus, we use machine learning methods, taking advantage of the near-optimal performance of some neural network decoders: multilayer perceptrons and convolutional neural networks (CNNs).  ...  For convolutional neural networks, we use the ResNet architecture, which allows us to implement very deep networks and results in better performance and scalability than the multilayer perceptron approach  ...  Dauphinais for useful discussions at the early stage of this research.  ... 
arXiv:2002.08666v2 fatcat:qvybmjjzwfflbpruplch5ivm2i

Resting State fMRI Functional Connectivity-Based Classification Using a Convolutional Neural Network Architecture

Regina J. Meszlényi, Krisztian Buza, Zoltán Vidnyánszky
2017 Frontiers in Neuroinformatics  
In this paper we describe a convolutional neural network architecture for functional connectome classification called connectomeconvolutional neural network (CCNN).  ...  We also show that the connectome-convolutional network is capable to combine information from diverse functional connectivity metrics and that models using a combination of different connectivity descriptors  ...  neural architectures at different noise and modification levels.  ... 
doi:10.3389/fninf.2017.00061 pmid:29089883 pmcid:PMC5651030 fatcat:iyurnydvanbyddbm347lfx2in4

DRNet: A Deep Neural Network With Multi-layer Residual Blocks Improves Image Denoising

Jiahong Zhang, Yonggui Zhu, Wenyi Li, Wenlong Fu, Lihong Cao
2021 IEEE Access  
Our study shows that the deep neural network with DC-ResBlocks, named DRNet, can achieve a very competitive result.  ...  DRNet performs well in denoising gray and color images with additive white Gaussian noise.  ...  These studies show that the deep convolution neural network can improve the image denoising effect. B. RESBLOCK IN DEEP NEURAL NETWORK Training a better network does not mean stacking more layers.  ... 
doi:10.1109/access.2021.3084951 fatcat:atxek5cihbhwbiihcv5uytj2eq

Self-Organized Operational Neural Networks for Severe Image Restoration Problems [article]

Junaid Malik, Serkan Kiranyaz, Moncef Gabbouj
2020 arXiv   pre-print
Discriminative learning based on convolutional neural networks (CNNs) aims to perform image restoration by learning from training examples of noisy-clean image pairs.  ...  However, the top-performing networks are generally composed of many convolutional layers and hundreds of neurons, with trainable parameters in excess of several millions.  ...  Figure 6 . 6 PSNR curves versus BP epochs for images restored from different noise models of all the networks. Table I . I Architectural details for the different networks.  ... 
arXiv:2008.12894v1 fatcat:wftcg2xr7zhj5b4io55xwkup7q
« Previous Showing results 1 — 15 out of 14,411 results