IEEE Journal of Selected Topics in Signal Processing最新文献_第8页

Energy Efficiency Optimization of IRS and UAV-Assisted Wireless Powered Edge Networks IRS和无人机辅助无线供电边缘网络的能效优化

IF 8.7 1区工程技术

IEEE Journal of Selected Topics in Signal Processing Pub Date : 2024-08-30 DOI: 10.1109/JSTSP.2024.3452501

Xiaojie Wang;Jiameng Li;Jun Wu;Lei Guo;Zhaolong Ning

{"title":"Energy Efficiency Optimization of IRS and UAV-Assisted Wireless Powered Edge Networks","authors":"Xiaojie Wang;Jiameng Li;Jun Wu;Lei Guo;Zhaolong Ning","doi":"10.1109/JSTSP.2024.3452501","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3452501","url":null,"abstract":"With the surge in the number of Internet of Things (IoT) devices and latency-sensitive services such as smart cities and smart factories, Next Generation Multiple Access (NGMA) technologies (e.g., Intelligent Reflecting Surface (IRS) and millimeter wave), which can efficiently process a large number of user accesses and low-latency services, have gained much attention. Among them, due to the ability to optimize wireless channels and improve data and energy transmission efficiency, IRS has been applied to Unmanned Aerial Vehicle (UAV)-assisted wireless powered edge networks. However, scheduling multi-dimensional resources in multi-UAVs, multi-IRSs and multi-devices coexistence scenarios always leads to a large number of highly coupled variables and complicated optimization problems. To address the above challenges, we propose a multi-agent Deep Reinforcement Learning (DRL)-based distributed scheduling algorithm for IRS and UAV-assisted wireless powered edge networks to jointly optimize charging time, phase shift matrices of IRSs, association scheduling of UAVs and UAV trajectories. First, to satisfy UAV time constraints and device energy consumption constraints, we formulate an energy efficiency maximization problem and represent it as a corresponding Markov Decision Process (MDP). Then, we propose a lightweight scheduling algorithm based on multi-agent DRL with value function decomposition. Finally, experiments show that the proposed algorithm has significant advantages in terms of algorithm convergence and system energy efficiency.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 7","pages":"1297-1310"},"PeriodicalIF":8.7,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993243","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}

引用次数: 0

Fractional Chirp Rate Based CSS Division Multiple Access Over LEO Satellite Internet-of-Things 基于分数啁啾率的低轨道卫星物联网上的CSS划分多址接入

IF 8.7 1区工程技术

IEEE Journal of Selected Topics in Signal Processing Pub Date : 2024-08-30 DOI: 10.1109/JSTSP.2024.3451290

Runxin Zhang;Jianpeng Ma;Shun Zhang;Octavia A. Dobre

{"title":"Fractional Chirp Rate Based CSS Division Multiple Access Over LEO Satellite Internet-of-Things","authors":"Runxin Zhang;Jianpeng Ma;Shun Zhang;Octavia A. Dobre","doi":"10.1109/JSTSP.2024.3451290","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3451290","url":null,"abstract":"Low earth orbit (LEO) satellites are bringing new opportunities for the integration between terrestrial Internet-of-Things (IoT) and satellite IoT. Due to its high robustness against large time delays and Doppler shifts, chirp spread spectrum (CSS) modulation, i.e., the key technology of the Long-Range (LoRa), is expected to empower the satellite link. However, the ALOHA protocol employed by LoRa will inevitably lead to collisions over the satellite channels. In this paper, we focus on the concurrent uplink transmission over the LEO satellite IoT, which is based on CSS. We carefully analyze the relationship between the chirp rate and its spreading factor (SF). Then, we propose the fractional chirp rate based CSS modulation, and support terrestrial users to achieve the non-orthogonal multiple access with the same SF, which ensures that the users possess the same noise immunity. We derive the bit error rate (BER) for both the synchronous and asynchronous scenarios. The performance of our scheme is tested by simulation. Results show that our scheme can achieve the multiple access while maintaining a satisfactory BER performance and is robust over the asynchronous scenario. Furthermore, we build a hardware system using the field-programmable gate array (FPGA) devices to validate the feasibility of this system.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 7","pages":"1281-1296"},"PeriodicalIF":8.7,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993244","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}

引用次数: 0

Beam-Sweeping Design for mmWave Massive Grant-Free Transmission 毫米波大规模无授权传输的波束扫描设计

IF 8.7 1区工程技术

IEEE Journal of Selected Topics in Signal Processing Pub Date : 2024-08-29 DOI: 10.1109/JSTSP.2024.3451706

Gangle Sun;Hongwei Hou;Yafei Wang;Wenjin Wang;Wei Xu;Shi Jin

{"title":"Beam-Sweeping Design for mmWave Massive Grant-Free Transmission","authors":"Gangle Sun;Hongwei Hou;Yafei Wang;Wenjin Wang;Wei Xu;Shi Jin","doi":"10.1109/JSTSP.2024.3451706","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3451706","url":null,"abstract":"To address the escalating demand for spectrum resources in emerging massive machine-type communication applications, it is promising to integrate massive grant-free transmission into millimeter-wave (mmWave) systems. As beam-sweeping schemes under hybrid beamforming architectures are commonly used to enhance signal power and extend coverage, this paper investigates an efficient beam-sweeping scheme for mmWave massive grant-free transmission under hybrid beamforming architectures. In this scheme, we propose a beam-sweeping design algorithm to optimize beamforming matrices, aiming to maximize spectral efficiency while ensuring quality of service (QoS) based on statistical information on uplink angles of arrival (AoAs). To address the intricate interdependence of beamforming matrices across different beam-sweeping slots, our solution begins with a two-stage genetic algorithm that pre-assigns users' access slots based on their uplink AoAs, decomposing the beamforming design problem into independent subproblems for each slot. Subsequently, a dual-layer beamforming design algorithm is proposed to solve these subproblems, optimizing beamforming matrices that enhance spectral efficiency and meet the QoS constraint. Numerous simulation results verify the effectiveness of the proposed beam-sweeping design algorithm in improving spectral efficiency and the capability to satisfy the required QoS.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 7","pages":"1249-1264"},"PeriodicalIF":8.7,"publicationDate":"2024-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993246","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}

引用次数: 0

Joint Fairness and Efficiency Optimization for CSMA/CA-Based Multi-User MIMO UAV Ad Hoc Networks 基于CSMA/ ca的多用户MIMO无人机自组网联合公平与效率优化

IF 8.7 1区工程技术

IEEE Journal of Selected Topics in Signal Processing Pub Date : 2024-08-27 DOI: 10.1109/JSTSP.2024.3435348

Jianrui Chen;Jingjing Wang;Jiaxing Wang;Lin Bai

{"title":"Joint Fairness and Efficiency Optimization for CSMA/CA-Based Multi-User MIMO UAV Ad Hoc Networks","authors":"Jianrui Chen;Jingjing Wang;Jiaxing Wang;Lin Bai","doi":"10.1109/JSTSP.2024.3435348","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3435348","url":null,"abstract":"Since conventional multiple access schemes, such as time division multiple access (TDMA), frequency division multiple access (FDMA), et al., cannot meet the requirements on flexibility, throughput and access delay in large-scale flying ad hoc networks (FANETs), we propose a fair and efficient CSMA/CA-based MAC protocol to facilitate concurrent uplink transmissions from different unmanned aerial vehicles (UAVs) by leveraging multiple-user MIMO (MU-MIMO). In this paper, we first propose the MIMO-based integrated sensing and backscatter communication model to achieve address resolution and also realize channel estimation by leveraging maximum likelihood estimation. Next, we propose an analytical model to characterize the saturation throughput and mean access delay of this CSMA/CA-based MAC protocol operating in an MU-MIMO FANET. Moreover, we derive the accurate expressions of saturation throughput and access delay under the proposed model. By means of the developed model, we evaluate the saturation throughput and access delay performance with respect to different network parameters, including the total payload, the number of UAVs and the number of UAV receiver's antennas. Numerical results indicate that our proposed protocol achieves superior throughput and decreased access delay.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 7","pages":"1311-1323"},"PeriodicalIF":8.7,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993325","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}

引用次数: 0

STAR-RIS-Aided UAV Communication for Next Generation Multiple Access With Resource Allocation 面向下一代多址资源分配的star - ris辅助无人机通信

IF 8.7 1区工程技术

IEEE Journal of Selected Topics in Signal Processing Pub Date : 2024-08-23 DOI: 10.1109/JSTSP.2024.3449124

Xiaoqi Zhang;Haijun Zhang;Lei Sun;Xianmei Wang;Keping Long;Victor C. M. Leung

{"title":"STAR-RIS-Aided UAV Communication for Next Generation Multiple Access With Resource Allocation","authors":"Xiaoqi Zhang;Haijun Zhang;Lei Sun;Xianmei Wang;Keping Long;Victor C. M. Leung","doi":"10.1109/JSTSP.2024.3449124","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3449124","url":null,"abstract":"Evolutionary non-orthogonal multiple access (NOMA) has emerged as a potential candidate for next-generation multiple access (NGMA). The signal processing research combined with NOMA and cutting-edge 6G technology reconfigurable intelligent surface (RIS) is attractive, but their compatibility is also a challenging research issue. In this paper, a full-coverage simultaneously transmitting and reflecting (STAR) RIS-enabled uplink NOMA communication model for unmanned aerial vehicle (UAV) is proposed to attain the joint design of computing and communication resources allocation. The aim of this work is to boost the sum-rate to the utmost degree while minimizing the system computing offloading latency. On the one hand, we jointly explore the computing offloading and NOMA uplink successive interference cancellation (SIC) decoding optimization problem in STAR-RIS assisted uplink signal propagation with energy splitting (ES) mode, solving for the computational offloading task size and decoding order. Alternatively, the beamforming and phase-shift of transmitting and reflecting is decoupled and optimized by hybrid whale-bat optimization algorithm (WBOA) and coordinate descent method (CDM), which gives a solution to minimize offloading latency. It is verified that the joint devised scheme may greatly contribute to the computing offloading capability and communication performance of the system.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 7","pages":"1222-1234"},"PeriodicalIF":8.7,"publicationDate":"2024-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142993247","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}

引用次数: 0

Optimizations of Neural Audio Coder Toward Perceptual Transparency

IF 8.7 1区工程技术

IEEE Journal of Selected Topics in Signal Processing Pub Date : 2024-08-22 DOI: 10.1109/JSTSP.2024.3437155

Joon Byun;Seungmin Shin;Seorim Hwang;Jongmo Sung;Seungkwon Beack;Youngcheol Park

{"title":"Optimizations of Neural Audio Coder Toward Perceptual Transparency","authors":"Joon Byun;Seungmin Shin;Seorim Hwang;Jongmo Sung;Seungkwon Beack;Youngcheol Park","doi":"10.1109/JSTSP.2024.3437155","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3437155","url":null,"abstract":"This paper presents comprehensive optimizations of a neural audio coder built upon a variational autoencoder (VAE) system integrated with an arithmetic coder. Our optimizations focus on two primary aspects: a novel loss function design and advanced entropy modeling of bottleneck latent embeddings. The loss function design incorporates parameters from a psychoacoustic model (PAM) into the frame-wise distortion measure, providing excellent perceptual quality. In addition, a multi-time scale discriminator is utilized to minimize distortions across adjacent frames, reducing artifacts at frame edges. Also, the coder is optimized considering three sophisticated entropy models within the latent domain: the Factorized Entropy Model (FEM), the Hyperprior Model (HPM), and the Joint Hierarchical Model (JHM). Notably, the JHM enhances context modeling across frames to effectively predict components influenced by long-term dependencies. To verify the optimization performance, we conducted extensive experiments using a dataset consisting of commercial movie clips and two additional public datasets. Objective metrics consistently demonstrated that our optimized loss function and latent modeling achieved superior performance across all test datasets compared to traditional codecs such as LAME-MP3 and FDK-AAC. Subjective assessments also indicated that our system could offer comparable or superior auditory quality to FDK-AAC.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 8","pages":"1531-1543"},"PeriodicalIF":8.7,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143184464","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}

引用次数: 0

S$^text{3}$Attention: Improving Long Sequence Attention With Smoothed Skeleton Sketching

IF 8.7 1区工程技术

IEEE Journal of Selected Topics in Signal Processing Pub Date : 2024-08-22 DOI: 10.1109/JSTSP.2024.3446173

Xue Wang;Tian Zhou;Jianqing Zhu;Jialin Liu;Kun Yuan;Tao Yao;Wotao Yin;Rong Jin;HanQin Cai

{"title":"S$^text{3}$Attention: Improving Long Sequence Attention With Smoothed Skeleton Sketching","authors":"Xue Wang;Tian Zhou;Jianqing Zhu;Jialin Liu;Kun Yuan;Tao Yao;Wotao Yin;Rong Jin;HanQin Cai","doi":"10.1109/JSTSP.2024.3446173","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3446173","url":null,"abstract":"Attention based models have achieved many remarkable breakthroughs in numerous applications. However, the quadratic complexity of Attention makes the vanilla Attention based models hard to apply to long sequence tasks. Various improved Attention structures are proposed to reduce the computation cost by inducing low rankness and approximating the whole sequence by sub-sequences. The most challenging part of those approaches is maintaining the proper balance between information preservation and computation reduction: the longer sub-sequences used, the better information is preserved, but at the price of introducing more noise and computational costs. In this paper, we propose a smoothed skeleton sketching based Attention structure, coined S<inline-formula><tex-math>$^{3}$</tex-math></inline-formula>Attention, which significantly improves upon the previous attempts to negotiate this trade-off. S<inline-formula><tex-math>$^{3}$</tex-math></inline-formula>Attention has two mechanisms to effectively minimize the impact of noise while keeping the linear complexity to the sequence length: a smoothing block to mix information over long sequences and a matrix sketching method that simultaneously selects columns and rows from the input matrix. We verify the effectiveness of S<inline-formula><tex-math>$^{3}$</tex-math></inline-formula>Attention both theoretically and empirically. Extensive studies over Long Range Arena (LRA) datasets and six time-series forecasting show that S<inline-formula><tex-math>$^{3}$</tex-math></inline-formula>Attention significantly outperforms both vanilla Attention and other state-of-the-art variants of Attention structures.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 6","pages":"985-996"},"PeriodicalIF":8.7,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106518","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}

引用次数: 0

NeaSource Localization and Beamforming in the Spherical Sector Harmonics Domain 球面扇形谐波域中的 NeaSource 定位和波束成形

IF 8.7 1区工程技术

IEEE Journal of Selected Topics in Signal Processing Pub Date : 2024-08-15 DOI: 10.1109/JSTSP.2024.3442469

Shekhar Kumar Yadav;S. R. M. Prasanna;Nithin V. George

{"title":"NeaSource Localization and Beamforming in the Spherical Sector Harmonics Domain","authors":"Shekhar Kumar Yadav;S. R. M. Prasanna;Nithin V. George","doi":"10.1109/JSTSP.2024.3442469","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3442469","url":null,"abstract":"Three-dimensional arrays can localize sources anywhere in the spatial domain without any ambiguity. Among these arrays, the spherical microphone array (SMA) has gained widespread usage in acoustic source localization and beamforming. However, SMAs are bulky, making them undesirable in applications with space and power constraints. To deal with this issue, arrays with microphones placed only in a sector of a sphere have been developed along with various techniques for localizing far-field sources in the spherical sector harmonics (S\u00002\u0000H) domain. This work addresses near-field acoustic localization and beamforming using a spherical sector microphone array. We first introduce a representation of spherical waves from a near-field point source in the S\u00002\u0000H domain using the orthonormal S\u00002\u0000H basis functions. Then, using the representation, we develop an array model for when a spherical sector array is placed in a wavefield created by multiple near-field sources in the S\u00002\u0000H domain. We highlight the advantages of the developed array model over the baseline near-field spatial domain array model. Using the developed array model, two algorithms are proposed for the joint estimation of the range, elevation and azimuth locations of near-field sources, namely NF-S\u00002\u0000H-MUSIC and NF-S\u00002\u0000H-MVDR. Further, a near-field beamforming algorithm capable of radial and angular filtering in the S\u00002\u0000H domain is also presented. Finally, we present the Cramer-Rao Bound (CRB) for range, elevation and azimuth estimation in the S\u00002\u0000H domain for near-field sources. The performances of the proposed algorithms are assessed using extensive near-field localization and beamforming simulations and an experiment.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"18 4","pages":"546-560"},"PeriodicalIF":8.7,"publicationDate":"2024-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142587556","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}

引用次数: 0

Time Scale Network: An Efficient Shallow Neural Network for Time Series Data in Biomedical Applications

IF 8.7 1区工程技术

IEEE Journal of Selected Topics in Signal Processing Pub Date : 2024-08-15 DOI: 10.1109/JSTSP.2024.3443659

Trevor Meyer;Camden Shultz;Najim Dehak;Laureano Moro-Velázquez;Pedro Irazoqui

{"title":"Time Scale Network: An Efficient Shallow Neural Network for Time Series Data in Biomedical Applications","authors":"Trevor Meyer;Camden Shultz;Najim Dehak;Laureano Moro-Velázquez;Pedro Irazoqui","doi":"10.1109/JSTSP.2024.3443659","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3443659","url":null,"abstract":"Time series data is often composed of information at multiple time scales, particularly in biomedical data. While numerous deep learning strategies exist to capture this information, many make networks larger, require more data, are more demanding to compute, and are difficult to interpret. This limits their usefulness in real-world settings facing even modest computational or data constraints and can further complicate their translation into real-time processing or edge device applicaitons. We present a minimal, computationally efficient Time Scale Network combining the translation and dilation sequence used in discrete wavelet transforms with traditional convolutional neural networks and back-propagation. The network simultaneously learns features at many time scales for sequence classification with significantly reduced parameters and operations. We demonstrate advantages in Atrial Dysfunction detection including: superior accuracy-per-parameter and accuracy-per-operation, fast training and inference speeds, and visualization and interpretation of learned patterns in atrial dysfunction detection on ECG signals. We also demonstrate impressive performance in seizure prediction using EEG signals, where our network isolated a few time scales that could be strategically selected to achieve 90.9% accuracy using only 1,133 active parameters and consistently converged on pulsatile waveform shapes. This method does not rest on any constraints or assumptions regarding signal content and could be leveraged in any area of time series analysis dealing with signals containing features at many time scales.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 1","pages":"129-139"},"PeriodicalIF":8.7,"publicationDate":"2024-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143512837","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}

引用次数: 0

Hypothesis Perturbation for Active Learning

IF 8.7 1区工程技术

IEEE Journal of Selected Topics in Signal Processing Pub Date : 2024-08-15 DOI: 10.1109/JSTSP.2024.3441311

Seong Jin Cho;Gwangsu Kim;Chang D. Yoo

{"title":"Hypothesis Perturbation for Active Learning","authors":"Seong Jin Cho;Gwangsu Kim;Chang D. Yoo","doi":"10.1109/JSTSP.2024.3441311","DOIUrl":"https://doi.org/10.1109/JSTSP.2024.3441311","url":null,"abstract":"This paper introduces a computationally efficient Query-by-Committee (QBC) algorithm specifically designed for deep active learning. The algorithm leverages the concept of <italic>hypothesis perturbation (HP) to construct the committee. The conventional QBC algorithms often incur high computational costs due to the independent training required for each committee member. In contrast, the HP constructs the committee by strategically sampling hypotheses around a given hypothesis, and efficiently identifies data points located near the decision boundary of the current hypothesis. To quantify uncertainty, the algorithm leverages a novel metric termed <italic>disagreement in hypothesis perturbation (DHP). DHP quantifies the disagreement in predictions between the given hypothesis and its perturbed hypotheses. This metric effectively identifies data points with high uncertainty, making them ideal candidates for active learning. The effectiveness of the proposed DHP-based active learning algorithm is empirically validated through extensive experimentation. The results demonstrate that the algorithm consistently achieves superior performance compared to other established algorithms across various datasets and deep network architectures considered in the study.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 1","pages":"115-128"},"PeriodicalIF":8.7,"publicationDate":"2024-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10637666","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143512971","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}

引用次数: 0