检索结果-内蒙古大学图书馆

Unimodal Training-Multimodal Prediction: Cross-modal Federated Learning with Hierarchical Aggregation

IEEE Transactions on Mobile Computing 2025年

作者： Zhang, Rongyu Chi, Xiaowei Zhang, Wenyi Liu, Guiliang Wang, Dan Wang, Fangxin Shenzhen China The Hong Kong Polytechnic University Department of Computing Hong Kong The Hong Kong University of Science and Technology Hong Kong Department of Electrical Engineering and Computer Science University of California Irvine United States Shenzhen China Shenzhen China Guangdong Provincial Key Laboratory of Future Networks of Intelligence China

Multimodal learning has significantly advanced the extraction of features from varied data sources, enhancing model performance. Federated learning (FL) complements this by enabling collaborative training while maintaining data privacy. The fusion of these two fields, multimodal federated learning, offers considerable promise. Yet, standard methods often incorrectly assume that each node in the FL network has a full complement of multimodal data, which is rare in real-world applications. In our study, we present a novel architecture designed to surmount these challenges, termed the Unimodal Training - Multimodal Prediction (UTMP) framework, positioned within the multimodal federated learning paradigm. Our proposed model, the HA-Fedformer, is a transformer-based model crafted to facilitate unimodal training on the client-side using exclusively unimodal datasets and to execute multimodal inference by synthesizing insights from multiple clients. Our HA-Fedformer model effectively handles non-IID data through a novel uncertainty-aware aggregation technique and layer-wise Markov Chain Monte Carlo sampling in local encoders. It also resolves misaligned language sequences via cross-modal decoder aggregation, capturing correlations between decoders trained on different modalities. Our comprehensive evaluations conducted on widely recognized sentiment analysis benchmarks demonstrate the superiority of the HA-Fedformer. The results show that our model achieves a substantial uplift in performance. © 2002-2012 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Integrated optical phased arrays: LiDAR, communications, augmented reality, and beyond

Integrated optical phased arrays: LiDAR, communications, aug...

引用

Photonic Networks and Devices, NETWORKS 2020

作者： Notaros, Jelena Notaros, Milica Raval, Manan Poulton, Christopher V. Byrd, Matthew J. Li, Nanxi Su, Zhan Magden, Emir Salih Timurdogan, Erman Dyer, Thomas Baiocco, Christopher Kim, Taehwan Bhargava, Pavan Stojanovic, Vladimir Watts, Michael R. Research Laboratory of Electronics Massachusetts Institute of Technology CambridgeMA02139 United States John A. Paulson School of Engineering and Applied Science Harvard University CambridgeMA02138 United States Colleges of Nanoscale Science and Engineering State University of New York Polytechnic Institute AlbanyNY12203 United States Department of Electrical Engineering and Computer Science University of California Berkeley BerkeleyCA94720 United States

Recent integrated optical phased array architectures, results, and applications will be reviewed, including beam steering for LiDAR and communications, near-field optical manipulation, and holographic displays for aug... 详细信息

ISBN: (纸本)9781943580798

关键词： Augmented reality

来源：评论

学校读者我要写书评

暂无评论

Low-damage electron beam lithography for nanostructures on Bi2Te3-class topological insulator thin films

arXiv

引用

arXiv 2023年

作者： Andersen, Molly P. Rodenbach, Linsey K. Rosen, Ilan T. Lin, Stanley C. Pan, Lei Zhang, Peng Tai, Lixuan Wang, Kang L. Kastner, Marc A. Goldhaber-Gordon, David Department of Materials Science and Engineering Stanford University StanfordCA94305 United States Stanford Institute for Materials and Energy Sciences SLAC National Accelerator Laboratory 2575 Sand Hill Road Menlo ParkCA94025 United States Department of Physics Stanford University StanfordCA94305 United States Department of Applied Physics Stanford University 348 Via Pueblo Mall StanfordCA94305 United States Stanford Nano Shared Facilities Stanford University StanfordCA94305 United States Department of Electrical and Computer Engineering Department of Physics and Astronomy University of California Los AngelesCA90095 United States Department of Physics Massachusetts Institute of Technology 77 Massachusetts Avenue CambridgeMA02139 United States

Nanostructured topological insulators (TIs) have the potential to impact a wide array of condensed matter physics topics, ranging from Majorana physics to spintronics. However, the most common TI materials, the Bi2Se3 family, are easily damaged during nanofabrication of devices. In this paper, we show that electron beam lithography performed with a 30 or 50 kV accelerating voltage – common for nanopatterning in academic facilities – damages both nonmagnetic TIs and their magnetically-doped counterparts at unacceptable levels. We additionally demonstrate that electron beam lithography with a 10 kV accelerating voltage produces minimal damage detectable through low-temperature electronic transport. Although reduced accelerating voltages present challenges in creating fine features, we show that with careful choice of processing parameters, particularly the resist, 100 nm features are reliably achievable. © 2023, CC BY.

关键词： Electron beam lithography

来源：评论

学校读者我要写书评

暂无评论

Nonreciprocal dynamics of ferrimagnetic bimerons

arXiv

引用

arXiv 2022年

作者： Shen, Laichuan Xia, Jing Chen, Zehan Li, Xiaoguang Zhang, Xichao Tretiakov, Oleg A. Shao, Qiming Zhao, Guoping Liu, Xiaoxi Ezawa, Motohiko Zhou, Yan School Of Science And Engineering The Chinese University Of Hong Kong Guangdong Shenzhen518172 China College Of Physics And Electronic Engineering Sichuan Normal University Chengdu610068 China Department Of Electronic And Computer Engineering The Hong Kong University Of Science And Technology Clear Water Bay Kowloon Hong Kong Department Of Physics The Hong Kong University Of Science And Technology Clear Water Bay Kowloon Hong Kong Center For Advanced Material Diagnostic Technology College Of Engineering Physics Shenzhen Technology University Shenzhen518118 China Department Of Electrical And Computer Engineering Shinshu University 4-17-1 Wakasato Nagano380-8553 Japan School Of Physics The University Of New South Wales Sydney2052 Australia Guangdong-Hong Kong-Macao Joint Laboratory For Intelligent Micro-Nano Optoelectronic Technology The Hong Kong University Of Science And Technology Hong Kong Hong Kong Department Of Applied Physics The University Of Tokyo 7-3-1 Hongo Tokyo113-8656 Japan

Magnetic bimerons are topologically nontrivial spin textures in in-plane easy-axis magnets, which can be used as particle-like information carriers. Here, we report a theoretical study on the nonreciprocal dynamics of asymmetrical ferrimagnetic (FiM) bimerons induced by spin currents. The FiM bimerons have the ability to move at a speed of kilometers per second and do not show the skyrmion Hall effect at the angular momentum compensation point. Our micromagnetic simulations and analytical results demonstrate that spin currents are able to induce the nonreciprocal transport and a drift motion of the FiM bimeron even if the system is at the angular momentum compensation point. By analyzing the current-induced effective fields, we find that the nonreciprocal transport is attributed to the asymmetry of the bimeron structure. Our results are useful for understanding the physics of bimerons in ferrimagnets and may provide guidelines for building bimeron-based spintronic devices. Copyright © 2022, The Authors. All rights reserved.

关键词： Ferrimagnetism

来源：评论

学校读者我要写书评

暂无评论

QoE-Driven transmission-aware cache placement and cooperative beamforming design in Cloud-RANs

QoE-Driven transmission-aware cache placement and cooperativ...

引用

作者： Sun, Ruijin Wang, Ying Cheng, Nan Lyu, Ling Zhang, Shan Zhou, Haibo Shen, Xuemin State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing100876 China School of Telecommunications Engineering State Key Laboratory of Integrated Services Network Xidian University Xi'an710071 China School of Information Science and Technology Dalian Maritime University Dalian116026 China School of Computer Science and Technology Beihang University Beijing100191 China School of Electronic Science and Engineering Nanjing University Nanjing210093 China Department of Electrical and Computer Engineering University of Waterloo WaterlooONN2L 3G1 Canada

Pre-caching popular videos at base stations (BSs) is a cost-effective way to significantly alleviate the backhaul pressure. With the video caching, the cache placement and the transmission strategy are intertwined with each other and jointly affect the system performance. Furthermore, the cache placement is updated in a much longer timescale than the transmission strategy. In this paper, the long-term transmission-aware cache placement and the short-term transmission strategy are designed to enhance the quality of experience (QoE) for the video streaming in cloud radio access networks (cloud-RANs). Specifically, consider a cache-enabled cloud-RAN, video contents are cached at BSs, and user requests are cooperatively satisfied by multiple BSs via the cooperative beamforming. To improve the weighted sum of users' QoE, the long-term transmission-aware caching problem in the caching stage and the short-term transmission problem in the delivery stage are respectively formulated, taking into account the backhaul capacity constraint, the transmission power constraint, and the storage size constraint. For the caching problem, the sample average approach is first used to approximate the long-term average QoE value. Then, cache placement strategies are devised in both centralized and distributed manner. For the transmission problem, the full-cooperative beamforming scheme is studied with the optimized cache placement, and an iterative algorithm is proposed. Simulation results show that our proposed transmission-aware cache placement and transmission strategies can achieve higher QoE performance than other cache placement and transmission strategies. © 1967-2012 IEEE.

关键词： Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Transfer learning approach to modeling multichannel gate-all-around nanosheet field-effect transistors under work function fluctuations

引用

engineering Applications of Artificial Intelligence 2025年 148卷

作者： Dash, Sagarika Li, Yiming Parallel and Scientific Computing Laboratory National Yang-Ming Chiao Tung University 1001 Ta-Hsueh Rd. Hsinchu300093 Taiwan Electrical Engineering and Computer Science International Graduate Program National Yang-Ming Chiao Tung University 1001 Ta-Hsueh Rd. Hsinchu300093 Taiwan Institute of Communications Engineering National Yang-Ming Chiao Tung University 1001 Ta-Hsueh Rd. Hsinchu300093 Taiwan Institute of Biomedical Engineering National Yang-Ming Chiao Tung University 1001 Ta-Hsueh Rd. Hsinchu300093 Taiwan Institute of Pioneer Semiconductor Innovation National Yang-Ming Chiao Tung University 1001 Ta-Hsueh Rd. Hsinchu300093 Taiwan Institute of Artificial Intelligence Innovation National Yang-Ming Chiao Tung University 1001 Ta-Hsueh Rd. Hsinchu300093 Taiwan Department of Electronics and Electrical Engineering National Yang-Ming Chiao Tung University 1001 Ta-Hsueh Rd. Hsinchu300093 Taiwan Department of Microelectronics National Yang-Ming Chiao Tung University 1001 Ta-Hsueh Rd. Hsinchu300093 Taiwan

Deep learning (DL) has significantly advanced various industries, including semiconductors, by providing sophisticated methods for analyzing emerging device data. Transfer learning (TL), a prominent DL topology, leverages knowledge from a pre-trained model to improve the performance of a specific target task. This study aims to apply TL to predict the impact of work function fluctuation (WKF) on the variability of threshold voltage (σVTH) in gate-all-around silicon nanosheet field-effect transistors (GAA Si NS FETs). We used TL to transfer knowledge from models trained on 1-channel (1ch) GAA Si NS FETs to more complex 4-channel (4ch) counterparts. The prediction accuracy is evaluated by using the mean square error(MSE), root mean square error (RMSE), and the mean absolute error(MAE) metrics. Our TL approach achieved an error rate of less than 1%, demonstrating high accuracy in predicting WKF effects on σ VTH of 4ch GAA Si NS FETs using only 500 samples. This study presents a pioneering application of TL using a one-dimensional (1D) convolutional neural network (CNN) long short-term memory (LSTM) topology in semiconductor device technology. The results effectively address the challenge of predicting σ VTH variations in advanced GAA Si NS FETs because of WKF, which eliminates the need for extensive semiconductor device simulations. © 2025 Elsevier Ltd

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Epilepsy Classification for Mining Deeper Relationships between EEG Channels based on GCN

Epilepsy Classification for Mining Deeper Relationships betw...

引用

2020 International Conference on computer Vision, Image and Deep Learning, CVIDL 2020

作者： Chen, Xin Zheng, Yuanjie Niu, Yi Li, Chengiiang Shandong Normal University School of Information Science and Engineering Jinan250358 China Key Lab of Intelligent Computing and Information Security Universities of Shandong Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Institute of Biomedical Sciences Jinan250358 China Department of Electrical Engineering Information Technology Shandong University of Science and Technology Jinan250031 China

ISBN: (纸本)9781728194813

Epilepsy is a brain disorder caused by abnormal discharges of neurons in brain. It is one of the most commonly studied disorders in neurology. The research of epilepsy electroencephalogram (EEG) has become a hot research topic. We find that in epilepsy EEG detection task, many previous methods focused on directly collecting the data of each channel, but these methods seldom analyse relationships between signals. Therefore, we propose the Epilepsy EEG Graph Convolutional Network EGCN, which makes full use of correlations between channels to deeply mine data information. We specifically design 5-layer graph convolutional network structure for classification of healthy and epileptic patients. The method is applied to public data set (Boon and CHB-MIT) to establish a reasonable classification model. And we compare it with some advanced algorithms. The experimental results show that the E-GCN method is superior to many existing methods in classification accuracy. In brief, the E-GCN method can be effectively used in classification and detection for epilepsy. This provides new ideas for colleagues, who study epilepsy EEG. In addition, this also provides richer experience for diagnosis of epilepsy. © 2020 IEEE.

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

From Missing Pieces to Masterpieces: Image Completion with Context-Adaptive Diffusion

arXiv

引用

arXiv 2025年

作者： Shamsolmoali, Pourya Zareapoor, Masoumeh Zhou, Huiyu Felsberg, Michael Tao, Dacheng Li, Xuelong Department of Computer Science University of York University of UK SCEE East China Normal University China Department of Electrical Engineering Shanghai Jiao Tong University China School of Computing and Mathematical Sciences University of Leicester United Kingdom Computer Vision Laboratory Linkoping University Sweden College of Computing and Data Science Nanyang Technological University Singapore of China Telecom China

Image completion is a challenging task, particularly when ensuring that generated content seamlessly integrates with existing parts of an image. While recent diffusion models have shown promise, they often struggle with maintaining coherence between known and unknown (missing) regions. This issue arises from the lack of explicit spatial and semantic alignment during the diffusion process, resulting in content that does not smoothly integrate with the original image. Additionally, diffusion models typically rely on global learned distributions rather than localized features, leading to inconsistencies between the generated and existing image parts. In this work, we propose ConFill, a novel framework that introduces a Context-Adaptive Discrepancy (CAD) model to ensure that intermediate distributions of known and unknown regions are closely aligned throughout the diffusion process. By incorporating CAD, our model progressively reduces discrepancies between generated and original images at each diffusion step, leading to contextually aligned completion. Moreover, ConFill uses a new Dynamic Sampling mechanism that adaptively increases the sampling rate in regions with high reconstruction complexity. This approach enables precise adjustments, enhancing detail and integration in restored areas. Extensive experiments demonstrate that ConFill outperforms current methods, setting a new benchmark in image completion. © 2025, CC BY-SA.

关键词： Photointerpretation

来源：评论

学校读者我要写书评

暂无评论

Regulating Crystallization Kinetics in High-Performance Perovskites Solar Modules Via Vapor Seed Layer engineering

SSRN

引用

SSRN 2024年

作者： Su, Zhenhuang Zhang, Juan He, Bingchen Huang, Jin Yang, Lin Ding, Yian Zhang, Wenjun Guo, Yukun Shi, Lei Abudula, Abuliti Du, Weizhi Hao, Xiaogang Ji, Xiaofei Yang, Liyou Guan, Guoqing Lu, Linfeng Gao, Xingyu The Interdisciplinary Research Center Shanghai Advanced Research Institute Chinese Academy of Sciences 99 Haike Road Zhangjiang Hi-Tech Park Pudong Shanghai201210 China Graduate School of Science and Technology Hirosaki University 3-Bunkyocho Hirosaki036-8561 Japan Jinneng Clean Energy Technology Ltd Shanxi Lvliang032100 China Shanghai Advanced Research Institute Chinese Academy of Sciences Shanghai China Hangzhou Zhongneng Photoelectricity Technology Co. Ltd. Hangzhou310018 China Department of Electrical Engineering and Computer Science University of Michigan Ann ArborMI48109 United States College of Chemical Engineering and Technology Taiyuan University of Technology Shanxi Taiyuan030024 China Energy Conversion Engineering Laboratory Institute of Regional Innovation Hirosaki University 3-Bunkyocho Hirosaki036-8561 Japan Huazhong University of Science and Technology China Zhejiang Sci-tech University China School of Chemistry and Chemical Engineering

Large-area fabrication of α-FAPbI₃ perovskite solar modules (PSMs) faces challenges in achieving uniform crystallinity and minimizing grain boundary defects. Herenin, a CsPbBr₃ seed layer was introduced onto the NiOx hole transport layer (HTL) via vacuum evaporation prior to perovskite film deposition using a vapor-blade coating method. Synchrotron-based in situ GIWAXS analysis revealed that the CsPbBr₃ seed layer effectively modulates the crystallization kinetics of PbI₂, facilitating the transition from δ-phase to α-phase perovskite and yielding films with superior crystallinity, grain size, and structural orientation. This seed layer also enhances the conductivity of NiOx, improves charge transport efficiency, and reduces recombination losses. As a result, large-area PSMs (active area: 61.56 cm²) incorporating the CsPbBr₃ seed layer achieved a power conversion efficiency (PCE) of 20.02%, compared to 17.62% for pristine devices. Additionally, these encapsulated modules exhibited excellent ambient stability, maintaining over 80% of their initial performance after 1100 hours under 60% relative humidity. This study highlights the potential of CsPbBr₃ seed layer engineering as a scalable and effective strategy for industrial production of high-efficiency, stable perovskite solar modules. © 2024, The Authors. All rights reserved.

关键词： Vacuum evaporation

来源：评论

学校读者我要写书评

暂无评论

Optimization of grant-free NOMA with multiple configured-grants for mURLLC

arXiv

引用

arXiv 2021年

作者： Liu, Yan Deng, Yansha Elkashlan, Maged Nallanathan, Arumugam Karagiannidis, George K. The Key Laboratory of Ministry of Education in Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications Nanjing210003 China School of Electronic Engineering and Computer Science Queen Mary University of London London United Kingdom Department of Engineering King’s College London London United Kingdom Department of Electrical and Computer Engineering Aristotle University of Thessaloniki Thessaloniki54 124 Greece

Massive Ultra-Reliable and Low-Latency Communications (mURLLC), which integrates URLLC with massive access, is emerging as a new and important service class in the next generation (6G) for time-sensitive traffics and has recently received tremendous research attention. However, realizing efficient, delay-bounded, and reliable communications for a massive number of user equipments (UEs) in mURLLC, is extremely challenging as it needs to simultaneously take into account the latency, reliability, and massive access requirements. To support these requirements, the third generation partnership project (3GPP) has introduced enhanced grant-free (GF) transmission in the uplink (UL), with multiple active configured-grants (CGs) for URLLC UEs. With multiple CGs (MCG) for UL, UE can choose any of these grants as soon as the data arrives. In addition, non-orthogonal multiple access (NOMA) has been proposed to synergize with GF transmission to mitigate the serious transmission delay and network congestion problems. In this paper, we develop a novel learning framework for MCG-GF-NOMA systems with bursty traffic. We first design the MCG-GF-NOMA model by characterizing each CG using the parameters: the number of contention-transmission units (CTUs), the starting slot of each CG within a subframe, and the number of repetitions of each CG. Based on the model, the latency and reliability performances are characterized. We then formulate the MCG-GF-NOMA resources configuration problem taking into account three constraints. Finally, we propose a Cooperative Multi-Agent based Double Deep Q-Network (CMA-DDQN) algorithm to balance the allocations of the channel resources among MCGs so as to maximize the number of successful transmissions under the latency constraint. Our results show that the MCG-GF-NOMA framework can simultaneously improve the low latency and high reliability performances in massive URLLC. Copyright © 2021, The Authors. All rights reserved.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：