IEEE Journal of Selected Topics in Signal Processing最新文献

{"title":"ESQFL: Digital Twin-Driven Explainable and Secured Quantum Federated Learning for Voltage Stability Assessment in Smart Grids","authors":"Chao Ren;Zhao Yang Dong;Han Yu;Minrui Xu;Zehui Xiong;Dusit Niyato","doi":"10.1109/JSTSP.2024.3485878","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3485878","url":null,"abstract":"Voltage stability remains a pivotal concern in power systems, especially with the integration of renewable energy sources and high-demand loads that use induction motors. Recent blackouts have highlighted the vulnerability of wind power systems to voltage stability assessment (VSA) threats. Traditional machine learning-based VSA methods, while efficient, often rely on centralized storage systems, making them susceptible to single-point failures. The incorporation of Digital Twins (DT), providing real-time virtual representations of physical power system components, offers transformative capabilities in prediction, analysis, and profit allocation within smart grids. This paper introduces an \u0000<underline>E</u>\u0000xplainable and \u0000<underline>S</u>\u0000ecured \u0000<underline>Q</u>\u0000uantum \u0000<underline>F</u>\u0000ederated \u0000<underline>L</u>\u0000earning (ESQFL) method for VSA, an innovative solution combining quantum techniques, differential privacy (DP), and Shapley value calculation. ESQFL, by leveraging the continuous insights from DT, emphasizes localized data analytics in a decentralized framework, integrating a Gaussian-based DP mechanism for enhanced data privacy and leveraging quantum teleportation for efficient Shapley value transmission. The paper systematically explores these concepts, compares the centralized and decentralized architectures, and provides comprehensive evaluations of ESQFL's efficacy on cross-out testing systems. The findings underscore ESQFL's potential as a pioneering solution in smart grid management, combining quantum computing with the advanced monitoring capabilities of DT for optimal VSA.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 5","pages":"964-978"},"PeriodicalIF":8.7,"publicationDate":"2024-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142938101","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Holographic Localization With Synthetic Reconfigurable Intelligent Surfaces","authors":"Ziyi Wang;Zhenyu Liu;Yuan Shen;Andrea Conti;Moe Z. Win","doi":"10.1109/JSTSP.2024.3435465","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3435465","url":null,"abstract":"Reconfigurable intelligent surfaces (RISs) are proposed to control complex wireless environments in next-generation networks. In particular, wideband RISs can play a key role in high-accuracy location awareness, which calls for models that consider the frequency-selectivity of metasurfaces. This paper presents a general signal model for wideband systems with RISs and establishes a Fisher information analysis to determine the theoretical limits of wideband localization with RISs. In addition, synthetic RISs are proposed to mitigate the multiplicative fading effect caused by the scattering property of RISs. Special scenarios including complete coupling and complete decoupling are further investigated. Results show that with the proposed models, a wideband RIS with a polynomial phase response per element provides more position information than those with more degrees of freedom (DOFs) in piecewise-constant phase response per element. Furthermore, velocity-induced information allows a dynamic RIS to provide more position information than a static RIS. Additionally, a dynamic RIS can be synthesized through multiple measurements to outperform a large one.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 4","pages":"603-618"},"PeriodicalIF":8.7,"publicationDate":"2024-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142587536","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DRED: Deep REDundancy Coding of Speech Using a Rate-Distortion-Optimized Variational Autoencoder","authors":"Jean-Marc Valin;Jan Büthe;Ahmed Mustafa;Michael Klingbeil","doi":"10.1109/JSTSP.2024.3482972","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3482972","url":null,"abstract":"Despite recent advancements in packet loss concealment (PLC) using deep learning techniques, packet loss remains a significant challenge in real-time speech communication. Redundancy has been used in the past to recover the missing information during losses. However, conventional redundancy techniques are limited in the maximum loss duration they can cover and are often unsuitable for burst packet loss. We propose a new approach based on a rate-distortion-optimized variational autoencoder (RDO-VAE), allowing us to optimize a deep speech compression algorithm for the task of encoding large amounts of redundancy at very low bitrate. The proposed Deep REDundancy (DRED) algorithm can transmit up to 50x redundancy using less than 32 kb/s. Results show that DRED outperforms the existing Opus codec redundancy. We also demonstrate its benefits when operating in the context of WebRTC.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 8","pages":"1441-1447"},"PeriodicalIF":8.7,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143184456","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MDCTNet: A Hybrid Approach to Neural Audio Coding","authors":"Lars Villemoes;Mark Vinton;Per Ekstrand;Lie Lu;Grant Davidson;Cong Zhou","doi":"10.1109/JSTSP.2024.3482721","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3482721","url":null,"abstract":"We describe and evaluate a hybrid neural audio coding system consisting of a perceptual audio encoder and a generative model, MDCTNet. By applying recurrent layers (RNNs) we capture correlations in both time and frequency directions in a perceptually weighted adaptive modified discrete cosine transform (MDCT) domain. By training MDCTNet on a diverse set of full-range monophonic audio signals at 48 kHz sampling, we achieve performance competitive with state-of-the-art audio coding at 24 kb/s variable bitrate (VBR). We also quantify the effect of the generative model-based decoding at lower and higher bitrates and discuss some caveats of the use of data driven signal reconstruction for the audio coding task.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 8","pages":"1506-1516"},"PeriodicalIF":8.7,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143184461","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrated Sensing and Communications for End-to-End Predictive Beamforming Design in Vehicle-to-Infrastructure Networks","authors":"Zihuan Wang;Vincent W.S. Wong;Robert Schober","doi":"10.1109/JSTSP.2024.3474254","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3474254","url":null,"abstract":"Integrated sensing and communications (ISAC) has emerged as a promising technology for predictive beamforming in vehicle-to-infrastructure (V2I) networks. Most of the existing works on ISAC assume each vehicle is equipped with a single antenna and use a two-phase scheme for predictive beamforming design. In the first phase, the reflected sensing signals at the roadside unit (RSU) are used to estimate the state parameters (e.g., angle, channel state information (CSI)) of the vehicles. In the second phase, the beamformer is predicted based on the estimated state parameters. The two-phase scheme suffers from the drawback that the estimation error in the first phase can impact the beamformer design in the second phase and may lead to a degradation in the achievable rate. In this work, we design predictive beamformers for both the RSU and vehicles in an end-to-end manner by using deep learning. We propose one-sided predictive beamforming (OSPB) and two-sided predictive beamforming (TSPB) schemes, where the beamformers for the vehicles are determined by the RSU and by the vehicles themselves, respectively. Both schemes directly predict the beamformers based on the reflected sensing signals via deep neural networks (DNNs). Compared with the existing two-phase schemes, the proposed schemes bypass the intermediate parameter estimation phase, thereby mitigating the impact of parameter estimation error. Our simulation results demonstrate the advantages of the proposed schemes over the two-phase baseline schemes in terms of achievable sum-rate.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 5","pages":"933-949"},"PeriodicalIF":8.7,"publicationDate":"2024-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142938139","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"HILCodec: High-Fidelity and Lightweight Neural Audio Codec","authors":"Sunghwan Ahn;Beom Jun Woo;Min Hyun Han;Chanyeong Moon;Nam Soo Kim","doi":"10.1109/JSTSP.2024.3469530","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3469530","url":null,"abstract":"The recent advancement of end-to-end neural audio codecs enables compressing audio at very low bitrates while reconstructing the output audio with high fidelity. Nonetheless, such improvements often come at the cost of increased model complexity. In this paper, we identify and address the problems of existing neural audio codecs. We show that the performance of the SEANet-based codec does not increase consistently as the network depth increases. We analyze the root cause of such a phenomenon and suggest a variance-constrained design. Also, we reveal various distortions in previous waveform domain discriminators and propose a novel distortion-free discriminator. The resulting model, <italic>HILCodec</i>, is a real-time streaming audio codec that demonstrates state-of-the-art quality across various bitrates and audio types.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 8","pages":"1517-1530"},"PeriodicalIF":8.7,"publicationDate":"2024-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143184462","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Lightweight SDENet Fusing Model-Based and Learned Features for Computational Histopathology","authors":"Rina Bao;Yunxin Zhao;Akhil Srivastava;Shellaine Frazier;Kannappan Palaniappan","doi":"10.1109/JSTSP.2024.3470312","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3470312","url":null,"abstract":"Model-based deep learning has the potential to significantly reduce the size of deep architectures while matching the competitive performance of much deeper and wider architectures. We demonstrate the advantage of combining model-based handcrafted features with learned features for AI-enabled computational pathology. Digital histopathology with whole slide image analysis using gigapixel-sized images and deep neural networks are being actively investigated for diagnosis and treatment, but require tens to hundreds of millions of learnable parameters (network weights). Additionally, using deep architectures effectively in medical applications, including pathology, has a number of challenges, including limited supervisory manual expert labels, a paucity of training data to cover clinical heterogeneity, many rare disease classes, complex anatomical structures and very large deep network architectures that have difficulty with domain adaptation along with generalization to new patient cohorts. We propose a lightweight squeeze, delineate, and excitation network (SDENet) deep learning architecture for pathology image cell and nuclei segmentation. SDENet is based on a novel hybrid network modular design, composed of a combination of model-based engineered or predefined filters for extracting salient information based on expert medical knowledge, with a learnable stack of convolution filters to capture structural relationships using complementary bottom-up data driven features. The proposed SDENetmodel-based approach learns rich feature representations of histopathology images, achieving a highly competitive performance on the MoNuSeg dataset for cell and nuclei segmentation with almost 90% fewer learnable parameters, and generalizes better to unseen image datasets, achieving about 20% higher accuracy on TNBC, compared to the widely used UNet architecture.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 6","pages":"1123-1137"},"PeriodicalIF":8.7,"publicationDate":"2024-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10697457","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106604","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"How to Train Your Unstable Looped Tensor Network","authors":"Anh-Huy Phan;Dmitri Ermilov;Nikolay Kozyrskiy;Igor Vorona;Konstantin Sobolev","doi":"10.1109/JSTSP.2024.3463480","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3463480","url":null,"abstract":"This paper addresses a substantial question of how to compress Deep Neural Networks with convolutional kernels modeled as looped tensor networks or Tensor Chain (TC) while it is known that such tensor network (TN) encounters severe numerical instability. We study the perturbation of this TN, provide an interpretation of instability in TC, propose novel methods to gain stability of the decomposition and keep the tensor network robust, and attain better approximation. Experimental results will confirm the superiority of the proposed methods in the compression of well-known convolutional neural networks, and TC decomposition under challenging scenarios.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 6","pages":"1036-1045"},"PeriodicalIF":8.7,"publicationDate":"2024-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106522","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Learning-Based Intermittent CSI Estimation With Adaptive Intervals in Integrated Sensing and Communication Systems","authors":"Jie Chen;Xianbin Wang","doi":"10.1109/JSTSP.2024.3468037","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3468037","url":null,"abstract":"Due to the distinct objectives and multipath utilization mechanisms between the communication and radar modules, the system design of integrated sensing and communication (ISAC) necessitates two types of channel state information (CSI), i.e., communication CSI representing the whole channel gain and phase shifts, and radar CSI exclusively focused on target mobility and position information. However, current ISAC systems apply an identical mechanism to estimate both types of CSI at the same predetermined estimation interval based on the worst case of dynamic environments, leading to significant overhead and compromised performances. Therefore, this paper proposes an intermittent communication and radar CSI estimation scheme with adaptive intervals for individual users/targets, where both types of CSI can be predicted using channel temporal correlations for cost reduction or re-estimated via signal transceiving for improved estimation accuracy. Specifically, we jointly optimize the binary CSI re-estimation/prediction decisions and transmit beamforming matrices for individual users/targets to maximize communication transmission rates and minimize radar tracking errors and costs in a multiple-input single-output (MISO) ISAC system. Unfortunately, this problem has causality issues because it requires comparing system performances under re-estimated CSI and predicted CSI during the optimization. However, the re-estimated CSI can only be obtained after completing the optimization. Additionally, the binary decision makes the joint design a mixed integer nonlinear programming (MINLP) problem, resulting in high complexity when using conventional optimization algorithms. Therefore, we propose a deep reinforcement online learning (DROL) framework that first implements an online deep neural network (DNN) to learn the binary CSI updating policy from the experiences. Given the learned policy, we propose an efficient algorithm to solve the remaining beamforming design problem. Finally, simulation results validate the effectiveness of the proposed algorithm.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 5","pages":"917-932"},"PeriodicalIF":8.7,"publicationDate":"2024-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142938103","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Designed Dithering Sign Activation for Binary Neural Networks","authors":"Brayan Monroy;Juan Estupiñan;Tatiana Gelvez-Barrera;Jorge Bacca;Henry Arguello","doi":"10.1109/JSTSP.2024.3467926","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3467926","url":null,"abstract":"Binary Neural Networks emerged as a cost-effective and energy-efficient solution for computer vision tasks by binarizing either network weights or activations. However, common binary activations, such as the Sign activation function, abruptly binarize the values with a single threshold, losing fine-grained details in the feature outputs. This work proposes an activation that applies multiple thresholds following dithering principles, shifting the Sign activation function for each pixel according to a spatially periodic threshold kernel. Unlike literature methods, the shifting is defined jointly for a set of adjacent pixels, taking advantage of spatial correlations. Experiments over the classification task using both grayscale and RGB datasets demonstrate the effectiveness of the designed dithering Sign activation function as an alternative activation for binary neural networks, without increasing the computational cost. Further, DeSign balances the preservation of details with the efficiency of binary operations.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 6","pages":"1100-1107"},"PeriodicalIF":8.7,"publicationDate":"2024-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106602","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}