检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Quan, Minh K. Nguyen, Dinh C. Nguyen, Van-Dinh Wijayasundara, Mayuri Setunge, Sujeeva Pathirana, Pubudu N. School of Engineering Deakin University Australia Department of Electrical and Computer Engineering University of Alabama Huntsville United States College of Engineering and Computer Science VinUniversity Vinhomes Ocean Park Hanoi100000 Viet Nam School of Engineering Royal Melbourne Institute of Technology University MelbourneVIC3000 Australia

Federated Learning is a promising approach for learning from user data while preserving data privacy. However, the high requirements of the model training process make it difficult for clients with limited memory or bandwidth to participate. To tackle this problem, Split Federated Learning is utilized, where clients upload their intermediate model training outcomes to a cloud server for collaborative server-client model training. This methodology facilitates resource-constrained clients' participation in model training but also increases the training time and communication overhead. To overcome these limitations, we propose a novel algorithm, called Hierarchical Split Federated Learning (HierSFL), that amalgamates models at the edge and cloud phases, presenting qualitative directives for determining the best aggregation time-frames to reduce computation and communication expenses. By implementing local differential privacy at the client and edge server levels, we enhance privacy during local model parameter updates. Our experiments using CIFAR-10 and MNIST datasets show that HierSFL outperforms standard FL approaches with better training accuracy, training time, and communication-computing trade-offs. HierSFL offers a promising solution to mobile edge computing's challenges, ultimately leading to faster content delivery and improved mobile service quality. Copyright © 2024, The Authors. All rights reserved.

关键词： Mobile edge computing

来源：评论

学校读者我要写书评

暂无评论

Design of a Low-Cost Wireless Charging Station Based on the Robotic Vision System

Design of a Low-Cost Wireless Charging Station Based on the ...

引用

International Symposium on High Capacity Optical Networks and Enabling Technologies, HONET

作者： Wen-Yu Cheng Wen-Chung Cheng Zhen Ni Erik D. Engeberg Xiangnan Zhong Department of Ocean and Mechanical Engineering Florida Atlantic University Boca Raton FL Department of Electrical Engineering & Computer Science Florida Atlantic University Boca Raton FL

Physical training of autonomous robotic systems are often limited by battery capacity since the intelligent training cycles have to be halted for battery recharge. Many researchers have focused on various solutions of power management for autonomous robots. However, these solutions are usually custom-fitted for the specific applications, and would be difficult to be generalized to other autonomous robotic systems. In this paper, we propose to develop a low-cost wireless charging system based on the on-board vision/camera sensor. This design is capable of autonomous docking for the robot chassis without further external supervision. Specifically, we use the QR codes to label the charging station and establish the robot's camera vision system to recognize the dock. The robot will drive to the station automatically. The low-cost charging station is also designed for wireless charging of the robot. We present the specific designs of charging dock, wireless connection, and battery charging experiments explicitly. This low-cost wireless charging system provides advancements toward a fully autonomous physical training infrastructure for robotic navigation experiments.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Novel Method for Tracking Neck Motions Using a Skin-Conformable Wireless Accelerometer: A Pilot Study

引用

Digital Biomarkers 2024年第1期8卷 40-51页

作者： Huang, Le Chun, Keum San Yu, Lian Lee, Jong Yoon Soetikno, Alan Chen, Hope Jeong, Hyoyoung Barrett, Joshua Martell, Knute Kang, Youn Patel, Alpesh A. Xu, Shuai Feinberg School of Medicine Northwestern University ChicagoIL United States Sibel Health NilesIL United States Electrical and Computer Engineering University of California Davis DavisCA United States Department of Orthopaedic Surgery Northwestern University Feinberg School of Medicine ChicagoIL United States Querrey Simpson Institute for Bioelectronics Northwestern University EvanstonIL United States Department of Ocean System Engineering Jeju National University Jeju Korea Republic of Departments of Orthopaedic Surgery and Neurosurgery Northwestern University Feinberg School of Medicine ChicagoIL United States Department of Dermatology Feinberg School of Medicine Northwestern University ChicagoIL United States

Introduction: Cervical spine disease is a leading cause of pain and disability. Degenerative conditions of the spine can result in neurologic compression of the cervical spinal cord or nerve roots and may be surgically treated with an anterior cervical discectomy and fusion (ACDF) in up to 137,000 people per year in the United States. A common sequelae of ACDF is reduced cervical range of motion (CROM) with patient-based complaints of stiffness and neck pain. Currently, tools for assessment of CROM are manual, subjective, and only intermittently utilized during doctor or physical therapy visits. We propose a skin-mountable acousto-mechanic sensor (ADvanced Acousto-Mechanic sensor;ADAM) as a tool for continuous neck motion monitoring in postoperative ACDF patients. We have developed and validated a machine learning neck motion classification algorithm to differentiate between eight neck motions (right/left rotation, right/left lateral bending, flexion, extension, retraction, protraction) in healthy normal subjects and patients. Methods: Sensor data from 12 healthy normal subjects and 5 patients were used to develop and validate a Convolutional Neural Network (CNN). Results: An average algorithm accuracy of 80.0 ± 3.8% was obtained for healthy normal subjects (94% for right rotation, 98% for left rotation, 65% for right lateral bending, 87% for left lateral bending, 89% for flexion, 77% for extension, 50% for retraction, 84% for protraction). An average accuracy of 67.5 ± 5.8% was obtained for patients. Discussion: ADAM, with our algorithm, may serve as a rehabilitation tool for neck motion monitoring in postoperative ACDF patients. Sensor-captured vital signs and other events (extubation, vocalization, physical therapy, walking) are potential metrics to be incorporated into our algorithm to offer more holistic monitoring of patients after cervical spine surgery. © 2024 The Author(s).

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Design of a Feeding System for Cage Aquaculture Based on IoT and AI Technology

Design of a Feeding System for Cage Aquaculture Based on IoT...

引用

2021 International Symposium on Intelligent Signal Processing and Communication Systems, ISPACS 2021

作者： Cheng, Chih-Yung Chang, Chin-Chun Lu, Hoang-Yang Cheng, Shyi-Chyi Nan, Fan-Hua Tang, Chi-Shih Hua, Zhen-Wei Lin, Wei-Yi Yu, Hsien-Jui National Taiwan Ocean University Department of Electrical Engineering Keelung Taiwan National Taiwan Ocean University Department of Computer Science and Engineering Keelung Taiwan National Taiwan Ocean University Department of Aquaculture Keelung Taiwan

ISBN: (纸本)9781665419512

This paper is about the design and integration of a novel offshore aquaculture feeding system based on IoT and AI techniques. An intelligent feeding system with fodder volume control and throwing accuracy can not only reduce waste and labor but also improve feeding efficiency. The proposed feeding system is connected to a cloud providing the volume to feed which is predicted according to various on-line environmental data such as temperature and water quality. Furthermore, the fodder throwing trajectory is detected and controlled through AI vision techniques. © 2021 IEEE.

关键词： Aquaculture

来源：评论

学校读者我要写书评

暂无评论

DX2CT: Diffusion Model for 3D CT Reconstruction from Bi or Mono-planar 2D X-ray(s)

DX2CT: Diffusion Model for 3D CT Reconstruction from Bi or M...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Yun Su Jeong Hye Bin Yoo Il Yong Chun Department of Electrical and Computer Engineering Sungkyunkwan University Republic of Korea Departments of Artificial Intelligence and Advanced Display Engineering Sungkyunkwan University Republic of Korea Center for Neuroscience Imaging Research Institute for Basic Science (IBS) Suwon Republic of Korea

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Computational tomography (CT) provides high-resolution medical imaging, but it can expose patients to high radiation. X-ray scanners have low radiation exposure, but their resolutions are low. This paper proposes a new conditional diffusion model, DX2CT, that reconstructs three-dimensional (3D) CT volumes from bi or mono-planar X-ray image(s). Proposed DX2CT consists of two key components: 1) modulating feature maps extracted from two-dimensional (2D) X-ray(s) with 3D positions of CT volume using a new transformer and 2) effectively using the modulated 3D position-aware feature maps as conditions of DX2CT. In particular, the proposed transformer can provide conditions with rich information of a target CT slice to the conditional diffusion model, enabling high-quality CT reconstruction. Our experiments with the bi or mono-planar X-ray(s) benchmark datasets show that proposed DX2CT outperforms several state-of-the-art methods. Our codes and model will be available at: https://***/intyeger/DX2CT.

关键词： Three-dimensional displays Computed tomography Two-dimensional displays Diffusion models Transformers Feature extraction Speech processing X-ray imaging Image reconstruction Biomedical imaging

来源：评论

学校读者我要写书评

暂无评论

Evaluating an Autonomous eDNA Sampler for Marine Environmental Monitoring: Short- and Long-Term Applications

Evaluating an Autonomous eDNA Sampler for Marine Environment...

引用

oceanS 2024 - Halifax, oceanS 2024

作者： Van Wyngaarden, Mallory Jeffery, Nicholas W. Horne, Edward Clemente-Carvalho, Rute B.G. Lemay, Matt Grundke, Iain Luy, Eddy Hendricks, Andre Sonnichsen, Colin Sieben, Vincent Laroche, Julie Beiko, Robert G. Furlong, Arnold Knox, Tom Pettitt-Wade, Harri Stanley, Ryan R.E. Bedford Institute of Oceanography Fisheries and Oceans Canada DartmouthNS Canada Tula Foundation Hakai Institute Campbell RiverBC Canada Dartmouth Ocean Technologies Inc DartmouthNS Canada Dalhousie University Department of Electrical & Computer Engineering HalifaxNS Canada Dalhousie University Department of Biology HalifaxNS Canada Dalhousie University Faculty of Computer Science HalifaxNS Canada

ISBN: (纸本)9798331540081

Monitoring biological diversity is essential for monitoring ecosystem health and evaluating conservation efforts. Environmental DNA metabarcoding has emerged as a powerful, scalable, and minimally invasive tool for assessing biodiversity. In this study, we evaluate the performance of an autonomous eDNA sampling platform deployed over a nine-week period and compare it with a traditional filter-at-sample (F AS) sampling protocol. Our results show that the autonomous sampling platform consistently captures and preserves DNA with comparable effectiveness to conventional in-situ filtering and preservation at -80°C. Using two eDNA markers, we found that fish communities (12S marker) identified by both sampling methods largely overlapped, while invertebrate detections (COI marker) differed between methods, likely due to differences in filter specifications. These findings demonstrate that the autonomous samplers worked effectively in comparison to traditional methods, highlighting their potential to expand the temporal and spatial coverage of eDNA-based biodiversity monitoring. The ability of these samplers to facilitate long-term and continuous sampling in challenging environments shows promise for advancing eDNA applications in diverse and remote settings. Further research is needed to assess their performance in deeper waters and over extended periods, particularly to evaluate eDNA preservation at ambient ocean temperatures. © 2024 IEEE.

关键词： Environmental monitoring

来源：评论

学校读者我要写书评

暂无评论

Sim-to-Real Brush Manipulation using Behavior Cloning and Reinforcement Learning

arXiv

引用

arXiv 2023年

作者： Jia, Biao Manocha, Dinesh The Department of Computer Science University of Maryland College ParkMD20740 United States The Departments of Computer Science and Electrical & Computer Engineering University of Maryland College ParkMD20740 United States

Developing proficient brush manipulation capabilities in real-world scenarios is a complex and challenging endeavor, with wide-ranging applications in fields such as art, robotics, and digital design. In this study, we introduce an approach designed to bridge the gap between simulated environments and real-world brush manipulation. Our framework leverages behavior cloning and reinforcement learning to train a painting agent, seamlessly integrating it into both virtual and real-world environments. Additionally, we employ a real painting environment featuring a robotic arm and brush, mirroring the MyPaint virtual environment. Our results underscore the agent’s effectiveness in acquiring policies for high-dimensional continuous action spaces, facilitating the smooth transfer of brush manipulation techniques from simulation to practical, real-world applications. © 2023, CC BY.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Evaluation of Low-Loss Polymer Switches for Multinuclear MRI/S*

Evaluation of Low-Loss Polymer Switches for Multinuclear MRI...

引用

Annual International Conference of the IEEE engineering in Medicine and Biology Society (EMBC)

作者： Edith Touchet-Valle Seelay Tasmim Taylor H. Ware Mary P. McDougall Department of Electrical and Computer Engineering Texas A&M University College Station TX USA Department of Biomedical Engineering Texas A&M University College Station TX USA Departments of Biomedical Engineering and Materials Science & Engineering Texas A&M University College Station TX USA Departments of Biomedical Engineering and Electrical and Computer Engineering Texas A&M University College Station TX USA

Implementation of multinuclear MRI/S as a diagnostic tool in clinical settings faces many challenges. One of those challenges is the development of highly sensitive multinuclear RF coils. Current multi-tuning techniques incorporate lossy components that impact the highest achievable SNR for at least one of the coil frequencies. As a result, optimization of multinuclear coil designs continues to be a priority for RF hardware engineers. To address this challenge, a new frequency switching technology that incorporates stimuli-responsive polymer materials was explored. Q measurements were used as a comparison metric between single-tuned, a standard switching network, and the proposed switching technology. The Q losses measured in the new switching method remained below 38% when compared to single-tuned coils. These results are consistent with low loss values reported using traditional switching networks. Furthermore, preliminary testing indicates that there is potential for improvement. These results establish the new technology as a promising alternative to traditional switching *** Relevance— A low loss multi-tuning technique for MRI radiofrequency coils has the potential of improving the study and diagnosis of disease.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Are Convolutional Neural Networks or Transformers more like human vision? 43

Are Convolutional Neural Networks or Transformers more like ...

引用

43rd Annual Meeting of the Cognitive science Society: Comparative Cognition: Animal Minds, CogSci 2021

作者： Tuli, Shikhar Dasgupta, Ishita Grant, Erin Griffiths, Thomas L. Department of Electrical and Computer Engineering Princeton University United States DeepMind New York United States Department of Electrical Engineering and Computer Sciences UC Berkeley United States Departments of Psychology and Computer Science Princeton University United States

Modern machine learning models for computer vision exceed humans in accuracy on specific visual recognition tasks, notably on datasets like ImageNet. However, high accuracy can be achieved in many ways. The particular decision function found by a machine learning system is determined not only by the data to which the system is exposed, but also the inductive biases of the model, which are typically harder to characterize. In this work, we follow a recent trend of in-depth behavioral analyses of neural network models that go beyond accuracy as an evaluation metric by looking at patterns of errors. Our focus is on comparing a suite of standard Convolutional Neural Networks (CNNs) and a recently-proposed attention-based network, the Vision Transformer (ViT), which relaxes the translation-invariance constraint of CNNs and therefore represents a model with a weaker set of inductive biases. Attention-based networks have previously been shown to achieve higher accuracy than CNNs on vision tasks, and we demonstrate, using new metrics for examining error consistency with more granularity, that their errors are also more consistent with those of humans. These results have implications both for building more human-like vision models, as well as for understanding visual object recognition in humans. © Cognitive science Society: Comparative Cognition: Animal Minds, CogSci *** rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Redundancy Is All You Need

arXiv

引用

arXiv 2024年

作者： Brakensiek, Joshua Guruswami, Venkatesan Department of Electrical Engineering and Computer Science University of California Berkeley United States Simons Institute for the Theory of Computing Departments of EECS and Mathematics University of California Berkeley United States

The seminal work of Benczúr and Karger demonstrated cut sparsifiers of near-linear size, with several applications throughout theoretical computer science. Subsequent extensions have yielded sparsifiers for hypergraph cuts and more recently linear codes over Abelian groups. A decade ago, Kogan and Krauthgamer asked about the sparsifiability of arbitrary constraint satisfaction problems (CSPs). For this question, a trivial lower bound is the size of a nonredundant CSP instance, which admits, for each constraint, an assignment satisfying only that constraint (so that no constraint can be dropped by the sparsifier). For instance, for graph cuts, spanning trees are non-redundant instances. Our main result is that redundant clauses are sufficient for sparsification: for any CSP predicate R, every unweighted instance of CSP(R) has a sparsifier of size at most its non-redundancy (up to polylog factors). For weighted instances, we similarly pin down the sparsifiability to the so-called chain length of the predicate. These results precisely determine the extent to which any CSP can be sparsified. A key technical ingredient in our work is a novel application of the entropy method from Gilmer's recent breakthrough on the union-closed sets conjecture. As an immediate consequence of our main theorem, a number of results in the non-redundancy literature immediately extend to CSP sparsification. We also contribute new techniques for understanding the non-redundancy of CSP predicates. In particular, we give an explicit family of predicates whose non-redundancy roughly corresponds to the structure of matching vector families in coding theory. By adapting methods from the matching vector codes literature, we are able to construct an explicit predicate whose non-redundancy lies between Ω(n1.5) and eO(n1.6), the first example with a provably non-integral exponent. © 2024, CC0.

关键词： Constraint satisfaction problems

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：