Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287777
Xinyue Liang, Alireza M. Javid, M. Skoglund, S. Chatterjee
We consider the problem of training a neural net over a decentralized scenario with a low communication over-head. The problem is addressed by adapting a recently proposed incremental learning approach, called ‘learning without forgetting’. While an incremental learning approach assumes data availability in a sequence, nodes of the decentralized scenario can not share data between them and there is no master node. Nodes can communicate information about model parameters among neighbors. Communication of model parameters is the key to adapt the ‘learning without forgetting’ approach to the decentralized scenario. We use random walk based communication to handle a highly limited communication resource.
{"title":"Learning without Forgetting for Decentralized Neural Nets with Low Communication Overhead","authors":"Xinyue Liang, Alireza M. Javid, M. Skoglund, S. Chatterjee","doi":"10.23919/Eusipco47968.2020.9287777","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287777","url":null,"abstract":"We consider the problem of training a neural net over a decentralized scenario with a low communication over-head. The problem is addressed by adapting a recently proposed incremental learning approach, called ‘learning without forgetting’. While an incremental learning approach assumes data availability in a sequence, nodes of the decentralized scenario can not share data between them and there is no master node. Nodes can communicate information about model parameters among neighbors. Communication of model parameters is the key to adapt the ‘learning without forgetting’ approach to the decentralized scenario. We use random walk based communication to handle a highly limited communication resource.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"44 1","pages":"2185-2189"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84582702","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287345
Haoyuan Cai, M. Kaloorazi, Jie Chen, Wei Chen, C. Richard
The generalized Hermitian eigendecomposition problem is ubiquitous in signal and machine learning applications. Considering the need of processing streaming data in practice and restrictions of existing methods, this paper is concerned with fast and efficient generalized eigenvectors tracking. We first present a computationally efficient algorithm based on randomization termed alternate-projections randomized eigenvalue decomposition (APR-EVD) to solve a standard eigenvalue problem. By exploiting rank-1 strategy, two online algorithms based on APR-EVD are developed for the dominant generalized eigenvectors extraction. Numerical examples show the practical applicability and efficacy of the proposed online algorithms.
{"title":"Online Dominant Generalized Eigenvectors Extraction Via A Randomized Method","authors":"Haoyuan Cai, M. Kaloorazi, Jie Chen, Wei Chen, C. Richard","doi":"10.23919/Eusipco47968.2020.9287345","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287345","url":null,"abstract":"The generalized Hermitian eigendecomposition problem is ubiquitous in signal and machine learning applications. Considering the need of processing streaming data in practice and restrictions of existing methods, this paper is concerned with fast and efficient generalized eigenvectors tracking. We first present a computationally efficient algorithm based on randomization termed alternate-projections randomized eigenvalue decomposition (APR-EVD) to solve a standard eigenvalue problem. By exploiting rank-1 strategy, two online algorithms based on APR-EVD are developed for the dominant generalized eigenvectors extraction. Numerical examples show the practical applicability and efficacy of the proposed online algorithms.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"41 1","pages":"2353-2357"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85128616","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287806
M. Hilton, Roxana Alexandru, P. Dragotti
We investigate the problem of reconstructing signals with finite rate of innovation from non-uniform samples obtained using an integrate-and-fire system. We assume that the signal is first filtered using the derivative of a hyperbolic secant as a sampling kernel. Timing information is then obtained using an integrator and a threshold detector. The reconstruction method we propose achieves perfect reconstruction of streams of K Diracs at arbitrary time locations, or equivalently piecewise constant signals with discontinuities at arbitrary time locations, using as few as 3K+1 non-uniform samples.
{"title":"Time Encoding Using the Hyperbolic Secant Kernel","authors":"M. Hilton, Roxana Alexandru, P. Dragotti","doi":"10.23919/Eusipco47968.2020.9287806","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287806","url":null,"abstract":"We investigate the problem of reconstructing signals with finite rate of innovation from non-uniform samples obtained using an integrate-and-fire system. We assume that the signal is first filtered using the derivative of a hyperbolic secant as a sampling kernel. Timing information is then obtained using an integrator and a threshold detector. The reconstruction method we propose achieves perfect reconstruction of streams of K Diracs at arbitrary time locations, or equivalently piecewise constant signals with discontinuities at arbitrary time locations, using as few as 3K+1 non-uniform samples.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"98 5 1","pages":"2304-2308"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83603598","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287335
Dimitris G. Chachlakis, Panos P. Markopoulos
We consider outlier-resistant Lp-quasi-norm (p ≤ 1) Principal-Component Analysis (Lp-PCA) of a D-by-N matrix. It was recently shown that Lp-PCA (p ≤ 1) admits an exact solution by means of combinatorial optimization with computational cost exponential in N. To date, apart from the exact solution to Lp-PCA (p ≤ 1), there exists no converging algorithm of lower cost that approximates its exact solution. In this work, we (i) propose a novel and converging algorithm that approximates the exact solution to Lp-PCA with significantly lower computational cost than that of the exact solver, (ii) conduct formal complexity and convergence analyses, and (iii) propose a multi-component solver based on subspace-deflation. Numerical studies on matrix reconstruction and medical-data classification illustrate the outlier resistance of Lp-PCA.
{"title":"Novel Algorithms for Lp-Quasi-Norm Principal-Component Analysis","authors":"Dimitris G. Chachlakis, Panos P. Markopoulos","doi":"10.23919/Eusipco47968.2020.9287335","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287335","url":null,"abstract":"We consider outlier-resistant Lp-quasi-norm (p ≤ 1) Principal-Component Analysis (Lp-PCA) of a D-by-N matrix. It was recently shown that Lp-PCA (p ≤ 1) admits an exact solution by means of combinatorial optimization with computational cost exponential in N. To date, apart from the exact solution to Lp-PCA (p ≤ 1), there exists no converging algorithm of lower cost that approximates its exact solution. In this work, we (i) propose a novel and converging algorithm that approximates the exact solution to Lp-PCA with significantly lower computational cost than that of the exact solver, (ii) conduct formal complexity and convergence analyses, and (iii) propose a multi-component solver based on subspace-deflation. Numerical studies on matrix reconstruction and medical-data classification illustrate the outlier resistance of Lp-PCA.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"83 1","pages":"1045-1049"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83776581","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287531
A. Mian, Elias Raninen, E. Ollila
In recent years, the use of Riemannian geometry has reportedly shown an increased performance for machine learning problems whose features lie in the symmetric positive definite (SPD) manifold. The present paper aims at reviewing several approaches based on this paradigm and provide a reproducible comparison of their output on a classic learning task of pedestrian detection. Notably, the robustness of these approaches to corrupted data will be assessed.
{"title":"A Comparative Study of Supervised Learning Algorithms for Symmetric Positive Definite Features","authors":"A. Mian, Elias Raninen, E. Ollila","doi":"10.23919/Eusipco47968.2020.9287531","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287531","url":null,"abstract":"In recent years, the use of Riemannian geometry has reportedly shown an increased performance for machine learning problems whose features lie in the symmetric positive definite (SPD) manifold. The present paper aims at reviewing several approaches based on this paradigm and provide a reproducible comparison of their output on a classic learning task of pedestrian detection. Notably, the robustness of these approaches to corrupted data will be assessed.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"3 1","pages":"950-954"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86648468","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287589
Cem Ates Musluoglu, A. Bertrand
The trace ratio optimization problem consists of maximizing a ratio between two trace operators and often appears in dimensionality reduction problems for denoising or discriminant analysis. In this paper, we propose a distributed and adaptive algorithm to solve the trace ratio optimization problem over network-wide covariance matrices, which capture the spatial correlation across sensors in a wireless sensor network. We focus on fully-connected network topologies, in which case the distributed algorithm reduces the communication bottleneck by only sharing a compressed version of the observed signals at each given node. Despite this compression, the algorithm can be shown to converge to the maximal trace ratio as if all nodes would have access to all signals in the network. We provide simulation results to demonstrate the convergence and optimality properties of the proposed algorithm.
{"title":"Distributed Trace Ratio Optimization in Fully-Connected Sensor Networks","authors":"Cem Ates Musluoglu, A. Bertrand","doi":"10.23919/Eusipco47968.2020.9287589","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287589","url":null,"abstract":"The trace ratio optimization problem consists of maximizing a ratio between two trace operators and often appears in dimensionality reduction problems for denoising or discriminant analysis. In this paper, we propose a distributed and adaptive algorithm to solve the trace ratio optimization problem over network-wide covariance matrices, which capture the spatial correlation across sensors in a wireless sensor network. We focus on fully-connected network topologies, in which case the distributed algorithm reduces the communication bottleneck by only sharing a compressed version of the observed signals at each given node. Despite this compression, the algorithm can be shown to converge to the maximal trace ratio as if all nodes would have access to all signals in the network. We provide simulation results to demonstrate the convergence and optimality properties of the proposed algorithm.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"20 1","pages":"1991-1995"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84284537","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287664
M. Laakso, Robin Rajamäki, R. Wichman, V. Koivunen
We introduce a modular and affordable coherent multichannel software-defined radio (SDR) receiver and demonstrate its performance by direction-of-arrival (DOA) estimation on signals collected from a 7 X 3 element uniform rectangular array antenna, comparing the results between the full and sparse arrays. Sparse sensor arrays can reach the resolution of a fully populated array with reduced number of elements, which relaxes the required structural complexity of e.g. antenna arrays. Moreover, sparse arrays facilitate significant cost reduction since fewer expensive RF-IF front ends are needed. Results from the collected data set are analyzed with Multiple Signal Classification (MUSIC) DOA estimator. Generally, the sparse array estimates agree with the full array.
{"title":"Phase-coherent multichannel SDR - Sparse array beamforming","authors":"M. Laakso, Robin Rajamäki, R. Wichman, V. Koivunen","doi":"10.23919/Eusipco47968.2020.9287664","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287664","url":null,"abstract":"We introduce a modular and affordable coherent multichannel software-defined radio (SDR) receiver and demonstrate its performance by direction-of-arrival (DOA) estimation on signals collected from a 7 X 3 element uniform rectangular array antenna, comparing the results between the full and sparse arrays. Sparse sensor arrays can reach the resolution of a fully populated array with reduced number of elements, which relaxes the required structural complexity of e.g. antenna arrays. Moreover, sparse arrays facilitate significant cost reduction since fewer expensive RF-IF front ends are needed. Results from the collected data set are analyzed with Multiple Signal Classification (MUSIC) DOA estimator. Generally, the sparse array estimates agree with the full array.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"31 1","pages":"1856-1860"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78319010","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287610
Bianca Iancu, E. Isufi
This work proposes a learning framework for distributed finite-time consensus with graph convolutional neural networks (GCNNs). Consensus is a central problem in distributed and adaptive optimisation, signal processing, and control. We leverage the link between finite-time consensus and graph filters, and between graph filters and GCNNs to study the potential of a readily distributed architecture for reaching consensus. We have found GCNNs outperform classical graph filters for distributed consensus and generalize better to unseen topologies such as distributed networks affected by link losses.
{"title":"Towards Finite-Time Consensus with Graph Convolutional Neural Networks","authors":"Bianca Iancu, E. Isufi","doi":"10.23919/Eusipco47968.2020.9287610","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287610","url":null,"abstract":"This work proposes a learning framework for distributed finite-time consensus with graph convolutional neural networks (GCNNs). Consensus is a central problem in distributed and adaptive optimisation, signal processing, and control. We leverage the link between finite-time consensus and graph filters, and between graph filters and GCNNs to study the potential of a readily distributed architecture for reaching consensus. We have found GCNNs outperform classical graph filters for distributed consensus and generalize better to unseen topologies such as distributed networks affected by link losses.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"22 1","pages":"2145-2149"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78504572","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287730
Alexander Bartler, Darius Hinderer, Bin Yang
Nowadays, the explainability of deep neural networks is an essential part of machine learning. In the last years, many methods were developed to visualize important regions of an input image for the decision of the deep neural network. Since almost all methods are designed for supervised trained models, we propose in this work a visualization technique for unsupervised trained autoencoders called Gradient-weighted Latent Activation Mapping (Grad-LAM). We adapt the idea of Grad-CAM and propose a novel weighting based on the knowledge of the autoencoder’s decoder. Our method will help to get insights into the highly nonlinear mapping of an input image to a latent space. We show that the visualization maps of Grad-LAM are meaningful on simple datasets like MNIST and the method is even applicable to real-world datasets like ImageNet.
{"title":"Grad-LAM: Visualization of Deep Neural Networks for Unsupervised Learning","authors":"Alexander Bartler, Darius Hinderer, Bin Yang","doi":"10.23919/Eusipco47968.2020.9287730","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287730","url":null,"abstract":"Nowadays, the explainability of deep neural networks is an essential part of machine learning. In the last years, many methods were developed to visualize important regions of an input image for the decision of the deep neural network. Since almost all methods are designed for supervised trained models, we propose in this work a visualization technique for unsupervised trained autoencoders called Gradient-weighted Latent Activation Mapping (Grad-LAM). We adapt the idea of Grad-CAM and propose a novel weighting based on the knowledge of the autoencoder’s decoder. Our method will help to get insights into the highly nonlinear mapping of an input image to a latent space. We show that the visualization maps of Grad-LAM are meaningful on simple datasets like MNIST and the method is even applicable to real-world datasets like ImageNet.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"68 1","pages":"1407-1411"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91347886","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287435
Tayebeh Lotfi Mahyari, R. Dansereau
One way to solve under-determined image decomposition is to use statistical information about the type of data to be decomposed. This information can be obtained by a deep learning where convolutional neural networks (CNN) are a subset recently used widely in image processing. In this paper, we have designed a two-stage CNN that takes cytology images of overlapped cervical cells and attempts to separate the cell images. In the first stage, we designed a CNN to segment overlapping cells. In the second stage, we designed a CNN that uses this segmentation and the original image to separate the regions. We implemented a CNN similar to U-Net for image segmentation and implemented a new network for the image separation. To train and test the proposed networks, we simulated 50000 cervical cell cytology images by overlaying individual images of real cervical cells using the Beer-Lambert law. Of these 50000 images, we used 49000 images for training and evaluated the method with 1000 test images. Results on these synthetic images give more than 97% segmentation accuracy and gives decomposition SSIM scores of more than 0.99 and PSNR score of more than 30 dB. Despite these positive results, the permutation problem that commonly effects signal separation occasionally occurred resulting in some cell structure mis-separation (for example, one cell given two nucleoli and the other given none). In addition, when the segmentation was poor from the first stage, the resulting separation was poor.
{"title":"Deep Learning Methods for Image Decomposition of Cervical Cells","authors":"Tayebeh Lotfi Mahyari, R. Dansereau","doi":"10.23919/Eusipco47968.2020.9287435","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287435","url":null,"abstract":"One way to solve under-determined image decomposition is to use statistical information about the type of data to be decomposed. This information can be obtained by a deep learning where convolutional neural networks (CNN) are a subset recently used widely in image processing. In this paper, we have designed a two-stage CNN that takes cytology images of overlapped cervical cells and attempts to separate the cell images. In the first stage, we designed a CNN to segment overlapping cells. In the second stage, we designed a CNN that uses this segmentation and the original image to separate the regions. We implemented a CNN similar to U-Net for image segmentation and implemented a new network for the image separation. To train and test the proposed networks, we simulated 50000 cervical cell cytology images by overlaying individual images of real cervical cells using the Beer-Lambert law. Of these 50000 images, we used 49000 images for training and evaluated the method with 1000 test images. Results on these synthetic images give more than 97% segmentation accuracy and gives decomposition SSIM scores of more than 0.99 and PSNR score of more than 30 dB. Despite these positive results, the permutation problem that commonly effects signal separation occasionally occurred resulting in some cell structure mis-separation (for example, one cell given two nucleoli and the other given none). In addition, when the segmentation was poor from the first stage, the resulting separation was poor.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"52 1","pages":"1110-1114"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90549014","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：