检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Buleje, Italo Siu, Vince S. Hsieh, Kuan Yu Hinds, Nigel Dang, Bing Bilal, Erhan Nguyen, Thanhnha Lee, Ellen E. Depp, Colin A. Rogers, Jeffrey L. Digital Health IBM T.J. Watson Research Center Yorktown HeightsNY United States Emerging Technology Engineering IBM T.J. Watson Research Center Yorktown Heights United States Department of Electrical and Computer Engineering College of Electrical and Computer Engineering National Yang Ming Chiao Tung University Hsinchu30010 Taiwan Institute of Biomedical Engineering College of Electrical and Computer Engineering National Yang Ming Chiao Tung University Hsinchu30010 Taiwan Department of Psychiatry University of California La Jolla San DiegoCA United States Sam and Rose Stein Institute for Research on Aging University of California La Jolla San DiegoCA United States USA IBM-UCSD Artificial Intelligence for Healthy Living Center La Jolla CA United States Veterans Affairs San Diego Healthcare System San DiegoCA United States

This paper presents a data-centric and security-focused data fabric designed for digital health applications. With the increasing interest in digital health research, there has been a surge in the volume of Internet of Things (IoT) data derived from smartphones, wearables, and ambient sensors. Managing this vast amount of data, encompassing diverse data types and varying time scales, is crucial. Moreover, compliance with regulatory and contractual obligations is essential. The proposed data fabric comprises an architecture and a toolkit that facilitate the integration of heterogeneous data sources, across different environments, to provide a unified view of the data in dashboards. Furthermore, the data fabric supports the development of reusable and configurable data integration components, which can be shared as open-source or inner-source software. These components are used to generate data pipelines that can be deployed and scheduled to run either in the cloud or on-premises. Additionally, we present the implementation of our data fabric in a home-based telemonitoring research project involving older adults, conducted in collaboration with the University of California, San Diego (UCSD). The study showcases the streamlined integration of data collected from various IoT sensors and mobile applications to create a unified view of older adults' health for further analysis and research. © 2023, CC BY-NC-ND.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Convolution-enhanced Evolving Attention Networks

arXiv

引用

arXiv 2022年

作者： Wang, Yujing Yang, Yaming Li, Zhuo Bai, Jiangang Zhang, Mingliang Li, Xiangtai Yu, Jing Zhang, Ce Huang, Gao Tong, Yunhai Key Laboratory of Machine Perception MOE School of Intelligence Science and Technology Peking University China Department of Computer Science & Technology Engineering Research Center of Microprocessor & System Peking University China Institute of Information Engineering Chinese Academy of Sciences China ETH Zürich Switzerland The Department of Automation Tsinghua University China

Attention-based neural networks, such as Transformers, have become ubiquitous in numerous applications, including computer vision, natural language processing, and time-series analysis. In all kinds of attention networks, the attention maps are crucial as they encode semantic dependencies between input tokens. However, most existing attention networks perform modeling or reasoning based on representations, wherein the attention maps of different layers are learned separately without explicit interactions. In this paper, we propose a novel and generic evolving attention mechanism, which directly models the evolution of inter-token relationships through a chain of residual convolutional modules. The major motivations are twofold. On the one hand, the attention maps in different layers share transferable knowledge, thus adding a residual connection can facilitate the information flow of inter-token relationships across layers. On the other hand, there is naturally an evolutionary trend among attention maps at different abstraction levels, so it is beneficial to exploit a dedicated convolution-based module to capture this process. Equipped with the proposed mechanism, the convolution-enhanced evolving attention networks achieve superior performance in various applications, including time-series representation, natural language understanding, machine translation, and image classification. Especially on time-series representation tasks, Evolving Attention-enhanced Dilated Convolutional (EA-DC-) Transformer outperforms state-of-the-art models significantly, achieving an average of 17% improvement compared to the best SOTA. To the best of our knowledge, this is the first work that explicitly models the layer-wise evolution of attention maps. Our implementation is available at https://***/pkuyym/EvolvingAttention. Copyright © 2022, The Authors. All rights reserved.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Survey on Deep Face Restoration: From Non-blind to Blind and Beyond

arXiv

引用

arXiv 2023年

作者： Li, Wenjie Wang, Mei Zhang, Kai Li, Juncheng Li, Xiaoming Zhang, Yuhang Gao, Guangwei Deng, Weihong Lin, Chia-Wen The Pattern Recognition and Intelligent System Laboratory School of Artificial Intelligence Beijing University of Posts and Telecommunications Beijing China The Computer Vision Lab ETH Zürich Zürich Switzerland The School of Communication and Information Engineering Shanghai University Shanghai China The Nanyang Technological University Singapore The Intelligent Visual Information Perception Laboratory Institute of Advanced Technology Nanjing University of Posts and Telecommunications Nanjing China The Department of Electrical Engineering National Tsing Hua University Hsinchu Taiwan

Face restoration (FR) is a specialized field within image restoration that aims to recover low-quality (LQ) face images into high-quality (HQ) face images. Recent advances in deep learning technology have led to significant progress in FR methods. In this paper, we begin by examining the prevalent factors responsible for real-world LQ images and introduce degradation techniques used to synthesize LQ images. We also discuss notable benchmarks commonly utilized in the field. Next, we categorize FR methods based on different tasks and explain their evolution over time. Furthermore, we explore the various facial priors commonly utilized in the restoration process and discuss strategies to enhance their effectiveness. In the experimental section, we thoroughly evaluate the performance of state-of-the-art FR methods across various tasks using a unified benchmark. We analyze their performance from different perspectives. Finally, we discuss the challenges faced in the field of FR and propose potential directions for future advancements. The open-source repository corresponding to this work can be found at https://***/24wenjie-li/Awesome-Face-Restoration. Copyright © 2023, The Authors. All rights reserved.

关键词： Restoration

来源：评论

学校读者我要写书评

暂无评论

Zero-Shot Audio Captioning Using Soft and Hard Prompts

arXiv

引用

arXiv 2024年

作者： Zhang, Yiming Xu, Xuenan Du, Ruoyi Liu, Haohe Dong, Yuan Tan, Zheng-Hua Wang, Wenwu Ma, Zhanyu The Pattern Recognition and Intelligent System Laboratory School of Artificial Intelligence Beijing University of Posts and Telecommunications Beijing100876 China The Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai200240 China The Department of Electronic Systems Aalborg University Aalborg9220 Denmark The Centre for Vision Speech and Signal Processing University of Surrey GuildfordGU2 7XH United Kingdom

In traditional audio captioning methods, a model is usually trained in a fully supervised manner using a human-annotated dataset containing audio-text pairs and then evaluated on the test sets from the same dataset. Such methods have two limitations. First, these methods are often data-hungry and require time-consuming and expensive human annotations to obtain audio-text pairs. Second, these models often suffer from performance degradation in cross-domain scenarios, i.e., when the input audio comes from a different domain than the training set, which, however, has received little attention. We propose an effective audio captioning method based on the contrastive language-audio pre-training (CLAP) model to address these issues. Our proposed method requires only textual data for training, enabling the model to generate text from the textual feature in the cross-modal semantic space. In the inference stage, the model generates the descriptive text for the given audio from the audio feature by leveraging the audio-text alignment from CLAP. We devise two strategies to mitigate the discrepancy between text and audio embeddings: a mixed-augmentation-based soft prompt and a retrieval-based acoustic-aware hard prompt. These approaches are designed to enhance the generalization performance of our proposed model, facilitating the model to generate captions more robustly and accurately. Extensive experiments on AudioCaps and Clotho benchmarks show the effectiveness of our proposed method, which outperforms other zero-shot audio captioning approaches for in-domain scenarios and outperforms the compared methods for cross-domain scenarios, underscoring the generalization ability of our method. © 2024, CC BY.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

A2G Channel Measurement and Characterization via TNN for UAV Multi-Scenario Communications

A2G Channel Measurement and Characterization via TNN for UAV...

引用

GLOBECOM 2022 - 2022 IEEE Global Communications Conference

作者： Kai Mao Qiuming Zhu Fuqiao Duan Yanheng Qiu Maozhong Song Wei Fan Yang Miao The Key Laboratory of Dynamic Cognitive System of Electromagnetic Spectrum Space College of Electronic and Information Engineering Nanjing University of Aeronautics and Astronautics Nanjing China The Radio Systems Faculty of Electrical Engineering Computer Science and Mathematics (EEMCS) University of Twente The Netherlands The Department of Electronic Systems Aalborg University Aalborg Denmark

ISBN: (纸本)9781665435413

Unmanned aerial vehicle (UAV) is considered as an important component for future communication networks. In this paper, an air-to-ground (A2G) channel sounder is designed and implemented for UAV communication channel measurement and characterization. The channel impulse response (CIR) extraction is implemented on a field programmable gate array (FPGA) to improve extraction efficiency. Based on the channel characteristics under measured (or baseline) scenarios, a transfer learning neural network (TNN) framework is also proposed to predict the channel characteristics of other unmeasured (or transferred) scenarios. In the proposed framework, the baseline matrices of neural network parameters are obtained from the measurement data of baseline scenarios. The ray tracing (RT) simulation data is only used to obtain the extrapolation matrices where we utilize imperfect digital map and do not require a highly accurate RT simulation. Then the neural network driven by the baseline and extrapolation matrices is used to predict the channel characteristics of transferred scenarios. To verify the proposed prediction method, the channel characteristics including path loss, K-factor, and root mean square delay spread of a near-urban scenario are firstly measured. Then, the corresponding channel characteristics of a transferred dense-urban scenario are predicted by the proposed TNN method and validated by the measurement data. It is shown that the predicted channel characteristics are well consistent with the measured ones.

关键词： Extrapolation Neural networks Transfer learning Prediction methods Ray tracing Autonomous aerial vehicles Loss measurement

来源：评论

学校读者我要写书评

暂无评论

Disentangling land model uncertainty via Matrix-based Ensemble Model Inter-comparison Platform(MEMIP)

引用

Ecological Processes 2022年第1期11卷 222-237页

作者： Cuijuan Liao Yizhao Chen Jingmeng Wang Yishuang Liang Yansong Huang Zhongyi Lin Xingjie Lu Yuanyuan Huang Feng Tao Danica Lombardozzi Almut Arneth Daniel SGoll Atul Jain Stephen Sitch Yanluan Lin Wei Xue Xiaomeng Huang Yiqi Luo Department of Earth System Science Ministry of Education Key Laboratory for Earth System ModelingInstitute for Global Change StudiesTsinghua UniversityBeijingChina Joint Innovation Center for Modern Forestry Studies College of Biology and the EnvironmentNanjing Forestry UniversityNanjingChina Department of Geography University of CambridgeCambridgeUK School of Life Science Nanjing UniversityNanjingChina Southern Marine Science and Engineering Guangdong Laboratory(Zhuhai) Guangdong Province Key Laboratory for Climate Change and Natural Disaster StudiesSchool of Atmospheric SciencesSun Yat-Sen UniversityGuangzhouChina CSIRO Oceans and Atmosphere Aspendale 3195Australia Terrestrial Sciences Section Climate and Global DynamicsNational Center for Atmospheric ResearchBoulderCO 80305USA Karlsruhe Institute of Technology Institute of Meteorology and Climate Research/Atmospheric Environmental ResearchGarmisch-PartenkirchenGermany UniversitéParis Saclay CEA-CNRS-UVSQLSCE/IPSLGif sur YvetteFrance Department of Atmospheric Sciences University of IllinoisUrbanaILUSA College of Life and Environmental Sciences University of ExeterExeterUK Department of Computer Science and Technology Tsinghua UniversityBeijingChina Center for Ecosystem Science and Society Department of Biological SciencesNorthern Arizona UniversityFlagstaffAZUSA

Background:Large uncertainty in modeling land carbon(C)uptake heavily impedes the accurate prediction of the global C *** the uncertainty sources among models is crucial for model improvement yet has been difficult due to multiple feedbacks within Earth system Models(ESMs).Here we present a Matrix-based Ensemble Model Inter-comparison Platform(MEMIP)under a unified model traceability framework to evaluate multiple soil organic carbon(SOC)*** the MEMIP,we analyzed how the vertically resolved soil biogeochemistry structure influences SOC prediction in two soil organic matter(SOM)*** comparing the model outputs from the C-only and CN modes,the SOC differences contributed by individual processes and N feedback between vegetation and soil were explicitly ***:Results showed that the multi-layer models with a vertically resolved structure predicted significantly higher SOC than the single layer models over the historical simulation(1900–2000).The SOC difference between the multi-layer models was remarkably higher than between the single-layer *** analysis indicated that over 80%of the SOC increase in the multi-layer models was contributed by the incorporation of depth-related processes,while SOC differences were similarly contributed by the processes and N feedback between models with the same soil depth ***:The output suggested that feedback is a non-negligible contributor to the inter-model difference of SOC prediction,especially between models with similar process *** analysis with TRENDY v7 and more extensive MEMIP outputs illustrated the potential important role of multi-layer structure to enlarge the current ensemble spread and the necessity of more detail model decomposition to fully disentangle inter-model *** stressed the importance of analyzing ensemble outputs from the fundamental model structures,and holding a holistic view in understanding the en

关键词： Soil organic carbon Inter-model comparison Uncertainty analysis Carbon-nitrogen coupling Vertical resolved soil biogeochemistry structure

来源：评论

学校读者我要写书评

暂无评论

Dynamic Hierarchical Skeleton-Based Network for Human Action Recognition in Video Content

SSRN

引用

SSRN 2024年

作者： Yang, Hualin Li, Xiying Chen, Calvin Yu-Chian Artificial Intelligence Medical Research Center School of Intelligent Systems Engineering Shenzhen Campus of Sun Yat-sen University Guangdong Shenzhen518107 China - Preferred Program Peking University Shenzhen Graduate School Guangdong Shenzhen518055 China School of Electronic and Computer Engineering Peking University Shenzhen Graduate School Guangdong Shenzhen518055 China Department of Medical Research China Medical University Hospital Taichung40447 Taiwan Department of Bioinformatics and Medical Engineering Asia University Taichung41354 Taiwan Guangdong Provincial Key Laboratory of Intelligent Transportation System Sun Yat-sen University Guangzhou China School of Intelligent Systems Engineering Sun Yat-sen University Guangdong Shenzhen518107 China

Human skeleton-based action recognition represents a pivotal field of study, capturing the intricate interplay between physical dynamics and intentional actions. Current research primarily focuses on extracting structural and temporal information from static skeleton-based graphs, but they grapple with a myriad of challenges. These include 1) an absence of hierarchical structure in encoding the skeleton-based graphs;2) A requirement for substantial prior knowledge to interpret the diverse spatial dynamics within singular action labels;3) An intricate task of representing the multifaceted temporal dynamics of individual actions. To counter this issue, we propose DHS-Net, an innovative dynamic hierarchical skeleton-based network in video content. The proposed DHS-Net incorporates the DHS Block as a central element, effectively facilitating information exchange through node concentration and diffusion across both structural and temporal dimensions. Through this architectural design, DHS-Net extracts hierarchical representations, enhancing the comprehensive understanding of spatio-temporal dynamics and marking a significant advancement in human action recognition. Our rigorous experimental evaluations further substantiate DHS-Net's supremacy, outperforming several state-of-the-art benchmarks in skeleton-based action recognition and achieving remarkable prediction accuracy. © 2024, The Authors. All rights reserved.

关键词： Dynamics

来源：评论

学校读者我要写书评

暂无评论

Efficient query evaluation techniques over large amount of distributed linked data

arXiv

引用

arXiv 2022年

作者： Kalogeros, Eleftherios Gergatsoulis, Manolis Damigos, Matthew Nomikos, Christos Laboratory on Digital Libraries and Electronic Publishing Department of Archives Library Science and Museology Ionian University Ioannou Theotoki 72 Corfu49100 Greece Department of Computer Science and Engineering University of Ioannina Greece P.O Box 1186 Ioannina45110 Greece

As RDF becomes more widely established and the amount of linked data is rapidly increasing, the efficient querying of large amount of data becomes a significant challenge. In this paper, we propose a family of algorithms for querying large amount of linked data in a distributed manner. These query evaluation algorithms are independent of the way the data is stored, as well as of the particular implementation of the query evaluation. We then use the MapReduce paradigm to present a distributed implementation of these algorithms and experimentally evaluate them, although the algorithms could be straightforwardly translated into other distributed processing frameworks. We also investigate and propose multiple query decomposition approaches of Basic Graph Patterns (subclass of SPARQL queries) that are used to improve the overall performance of the distributed query answering. A deep analysis of the effectiveness of these decomposition algorithms is also provided. Copyright © 2022, The Authors. All rights reserved.

关键词： Big data

来源：评论

学校读者我要写书评

暂无评论

Tolerating Device-to-Device Variation for Memristive Crossbar-Based Neuromorphic Computing systems: A New Bayesian Perspective

Tolerating Device-to-Device Variation for Memristive Crossba...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Yang Xiao Qi Xu Bo Yuan Research Institute of Trustworthy Autonomous System (RITAS) Southern University of Science and Technology (SUSTech) Shenzhen China Department of Computer Science and Engineering Guangdong Provincial Key Laboratory of Brain-inspired Intelligent Computation Southern University of Science and Technology (SUSTech) Shenzhen China School of Microelectronics University of Science and Technology of China (USTC)

Memristive crossbar-based architecture provides an energy-efficient platform to accelerate neural networks (NNs) thanks to its Processing-in-Memory (PIM) nature. However, the device-to-device variation (DDV), which is typically modeled as Lognormal distribution, deviates the programmed weights from their target values, resulting in significant performance degradation. This paper proposes a new Bayesian Neural Network (BNN) approach to enhance the robustness of weights against DDV. Instead of using the widely-used Gaussian variational posterior in conventional BNNs, our approach adopts a DDV-specific variational posterior distribution, i.e., Lognormal distribution. Accordingly, in the new BNN approach, the prior distribution is modified to keep consistent with the posterior distribution to avoid expensive Monte Carlo simulations. Furthermore, the mean of the prior distribution is dynamically adjusted in accordance with the mean of the Lognormal variational posterior distribution for better convergence and accuracy. Compared with the state-of-the-art approaches, experimental results show that the proposed new BNN approach can significantly boost the inference accuracy with the consideration of DDV on several well-known datasets and modern NN architectures. For example, the inference accuracy can be improved from 18% to 74% in the scenario of ResNet-18 on CIFAR-10 even under large variations.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-Step Ahead Wind Power Forecasting Based on Dual-Attention Mechanism

SSRN

引用

SSRN 2022年

作者： Aslam, Muhammad Kim, Jun-Sung Jung, Jaesung Department of Artificial Intelligence Sejong University Seoul05006 Korea Republic of Digital Solution Laboratory Korea Electric Power Research Institute Daejeon34056 Korea Republic of Department of Electrical and Computer Engineering Ajou University Gyeonggi-do Suwon16499 Korea Republic of

Accurate forecasting is essential for the economic benefits and efficient operation of intermittent wind power systems. Multi-step ahead wind power forecasting provides multiple benefits in the planning and operation of the power systems. This study proposes a deep learning model based on a dual-attention mechanism for multi-step ahead wind power forecasting. Both the attention mechanisms are applied over the encoder-decoder based sequence-to-sequence model consisting of long short-term memory (LSTM) blocks. The Bayesian optimization algorithm is applied to the proposed model to obtain the optimal combination of hyper-parameters. The proposed model has been compared with the persistence model, and state-of-the-art models such as simple LSTM, LSTM-attention, ensemble model, and neural basis expansion analysis for time series model to evaluate its effectiveness. Furthermore, the performance of the attention mechanism with respect to impactful input features, such as wind speed and air pressure, was analyzed. The forecasting skill score of the proposed model was the highest among all other models in comparison, which indicates the effectiveness of the proposed model. Similarly, the proposed model outperformed the traditional methods in terms of other evaluating criteria, including mean absolute error, and root mean square error among others, hence, proving its efficacy. © 2022, The Authors. All rights reserved.

关键词： Long short-term memory

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：