检索结果-内蒙古大学图书馆

Overview of Essential Components in deep learning reference-based super resolution methods

Procedia Computer Science 2024年 242卷 1243-1248页

作者： Jiayu Xue Junjie Liu Yong Shi School of Computer Science and Technology University of Chinese Academy of Sciences Beijing 101408 China Research Center on Fictitious Economy & Data Science Chinese Academy of Sciences Beijing 100190 China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing 100190 China Sino-Danish College University of Chinese Academy of Sciences Beijing 100049 China School of Economics and Management University of Chinese Academy of Sciences Beijing 100190 China College of Information Science and Technology University of Nebraska at Omaha Omaha NE 68182 USA

Reference-based super resolution (RefSR) aims to recover the lost details in a low-resolution image and generate a high-resolution result, guided by a high-resolution reference image with similar contents or textures. In contrast to the traditional single-image super-resolution, which focuses on the intrinsic properties of the single low-resolution image, the challenge of RefSR lies in matching and aggregating highly-related but misaligned reference images with low-resolution images. Several effective but complex designs have been proposed to address this challenge, which poses difficulties in implementing RefSR in real-world applications. In order to better understand the working mechanism of RefSR and design a more efficient and lightweight architecture, we provide a review about the essential components of the existing deep learning-based RefSR. We decompose and classify the common pipeline into four submodules according to the functionalities. Then, we summarize and describe the implementation details of the commonly-adopted approaches in each submodule. Finally, we discuss the challenges and promising research directions.

关键词： Reference-based super resolution Deep learning Feature matching Feature alignment

来源：评论

学校读者我要写书评

暂无评论

An ES-based model with contemporaneous and temporal aggregation for forecasting intermittent and lumpy retail demand

引用

Procedia Computer Science 2024年 242卷 264-271页

作者： Erjiang E Xin Tian Ye Tao Huiqiang Sun School of Management Guangxi Minzu University Nanning 530006 China Research Center on Fictitious Economy and Data Science Chinese Academy of Sciences Beijing 100190 China School of Economics and Management University of Chinese Academy of Sciences Beijing 100190 China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy Sciences Beijing 100190 China Beijing Haolinju CVS Co. Ltd. Beijing 100018 China Shenghe (Beijing) Technology Development Co. Ltd. Beijing 101399 China

The demand for products in the retail industry is often characterized by intermittent and volatile patterns. Specifically, at the SKU level, the demand exhibits intermittent and lumpy behavior, which presents challenges for accurate forecasting due to demand fluctuations and interval uncertainty. This study proposes an exponential smoothing (ES)-based model, called ES-C-T, that incorporates contemporaneous and temporal aggregation to forecast intermittent and lumpy retail demand. The ES-C-T model consists of three stages. Firstly, the ES model predicts the sales of each product by aggregating contemporaneously across all stores, and the resulting predictions are allocated to the SKU level using a weight vector. Secondly, the ES model forecasts the sales of each SKU by aggregating temporally within each store. Finally, the ES-C-T model derives a weight coefficient based on the predictions from the first and second stages, adjusts the initial predictions, and utilizes this as the final forecast. To validate the effectiveness of the ES-C-T model, sales data from a large Chinese convenience store chain are employed. The results demonstrate that the proposed ES-C-T model outperforms the benchmark model in terms of MAE, RMSE, and WRMSSE, effectively predicting retail sales for intermittent and lumpy demand.

关键词： forecasting retailing intermittent demand lumpy demand aggregation

来源：评论

学校读者我要写书评

暂无评论

A Routing Optimization in Wireless Sensor Networks Based on Reverse Elite Sparrow Search Algorithm

Journal of Network Intelligence

引用

Journal of Network Intelligence 2021年第4期6卷 763-775页

作者： Ning, Chun-Yuan Liao, Lyu-Chao Nguyen, Trong-The Huang, Kuan-Chun Pan, Jeng-Shyang Fujian Provincial Key Laboratory of Big Data Mining and Applications Fujian University of Technology Fuzhou350118 China Multimedia Communications Lab University of Information Technology VNU-HCM Viet Nam Department of Industrial and Systems Engineering Chung Yuan Christian University Taoyuan320314 Taiwan College of Computer Science and Engineering Shandong University of Science and Technology Qingdao266590 China

This study suggests a solution to the routing optimization insufficiency based on the reverse elite sparrow search algorithm (RESSA) for energy balancing and network lifetime prolonging in wireless sensor networks (WSN). The node’s consumption and the remaining energy priority to balance consumption in the data forwarding to the cluster head are considered the optimal path evaluation function. The obtained results of the experimental the suggested scheme are compared with the other methods show that the proposed scheme increases data forwarding round numbers and consumes less energy, the remaining energy more, and higher efficiency of node energy consumption. © The Authors.

关键词： Energy utilization

来源：评论

学校读者我要写书评

暂无评论

A Secure data Deduplication Scheme Based on Differential Privacy

A Secure Data Deduplication Scheme Based on Differential Pri...

引用

International Conference on Parallel and Distributed Systems (ICPADS)

作者： Jun Ren Zhiqiang Yao Jinbo Xiong Yuanyuan Zhang Ayong Ye Faculty of Software Fujian Normal University Fuzhou China Fujian Engineering Research Center of Public Service Big Data Mining and Application Furhou China Fujian Normal University Fuzhou Fujian CN Fujian Provincial Key Laboratory of Network Security and Cryptology FJNU Fuzhou China

ISBN: (纸本)9781509053827

In cloud computing environment, especially in Big data era, adversary may use data deduplication service supported by the cloud service provider as a side channel to eavesdrop users' privacy or sensitive information. In order to tackle this serious issue, in this paper, we propose a secure data deduplication scheme based on differential privacy. The highlights of the proposed scheme lie in constructing a hybrid cloud framework, using convergent encryption algorithm to encrypt original files, and introducing differential privacy mechanism to resist against the side channel attack. Performance evaluation shows that our scheme is able to effectively save network bandwidth and disk storage space during the processes of data deduplication. Meanwhile, security analysis indicates that our scheme can resist against the side channel attack and related files attack, and prevent the disclosure of privacy information.

关键词： Conferences

来源：评论

学校读者我要写书评

暂无评论

An improving data compression capability in sensor node to support sensorml-compatible for internet-of-things

Journal of Network Intelligence

引用

Journal of Network Intelligence 2018年第2期3卷 74-90页

作者： Nguyen, Trong-The Dao, Thi-Kien Pan, Jeng-Shyang Horng, Mong-Fong Shieh, Chin-Shiuh Department of Information Technology Haiphong Private University No. 35 Danlap Rd. Hai-Phong Viet Nam Fujian Provincial Key Laboratory of Big Data Mining and Applications Fujian University of Technology Fujian China College of Information Science and Engineering Fujian University of Technology Fujian China Department of Electronics Engineering National Kaohsiung University of Applied Sciences Taiwan

Limited bandwidth and lacked compatibility information of the sensor nodes in Wireless Sensor Network (WSN) are critical issues in implementing WSN applica-tions. This paper proposes a method of improvement to data compression capability to support SensorML interface for information exchange in the sensor nodes. The delivery of the packets with XML format of all nodes in WSN could cause the traffic load in-creases. This paper proposes a method of improvement of data compression capability for exchanging data in the sensorML for the Internet of Things (IoT). The delivery of XML formatted packets of all nodes in Wireless sensor networks (WSN) could cause the traffic load increases. The proposed method suggests data compression condensing the packets in the node and analyzes the relationship between the parameters and the performance. The settings of sliding window size and comparing length are factors to affect the performance of network traffic load significantly. The experimental results show that, with a format of 256 bytes and LZSS compression, the transmission performance is improved up to 48.8%, in comparison with the original approach. Besides, the proposed method can deliver up to 204% of information than the other works. © 2018, Taiwan Ubiquitous Information. All rights reserved.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Fuzzy-NMS: Improving 3D Object Detection With Fuzzy Classification in NMS

IEEE Transactions on Intelligent Vehicles

引用

IEEE Transactions on Intelligent Vehicles 2024年 1-15页

作者： Wang, Li Zhang, Xinyu Zhao, Fachuan Wu, Chuze Wang, Yichen Song, Ziying Yang, Lei Xu, Bin Li, Jun Ge, Shuzhi Sam School of Machanical Engineering Beijing Institute of Technology Beijing China School of Vehicle and Mobility Tsinghua University Beijing China School of Computer and Information Technology Beijing Key Lab of Traffic Data Analysis and Mining Beijing Jiaotong University Beijing China Department of Electrical and Computer Engineering Interactive Digital Media Institute Social Robotics Laboratory National University of Singapore Singapore

Non-maximum suppression (NMS) is an essential post-processing module in many 3D object detection frameworks to remove overlapping candidate bounding boxes. However, an overreliance on classification scores and difficulties in determining appropriate thresholds can affect the resulting accuracy directly. To address these issues, we introduce fuzzy learning into NMS and propose a novel generalized Fuzzy-NMS module to achieve finer candidate bounding box filtering. The proposed Fuzzy-NMS module combines the volume and clustering density of candidate bounding boxes, refining them with a fuzzy classification method and optimizing the appropriate suppression thresholds to reduce uncertainty in the NMS process. Adequate validation experiments use the mainstream KITTI and large-scale Waymo 3D object detection benchmarks. The results of these tests demonstrate the proposed Fuzzy-NMS module can improve the accuracy of numerous recently NMS-based detectors significantly, including PointPillars, PV-RCNN, and IA-SSD, etc. This effect is particularly evident for small objects such as pedestrians and bicycles. As a plug-and-play module, Fuzzy-NMS does not need to be retrained and produces no obvious increases in inference time. IEEE

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

An overview on visualization of ontology alignment and ontology entity 5th

An overview on visualization of ontology alignment and ontol...

引用

5th Euro-China Conference on Intelligent data Analysis and Applications, ECC 2018

作者： Chen, Jie Xue, Xingsi Huang, Lili Ren, Aihong College of Information Science and Engineering Fujian University of Technology FuzhouFujian350118 China Intelligent Information Processing Research Center Fujian University of Technology FuzhouFujian350118 China Fujian Provincial Key Laboratory of Big Data Mining and Applications FuzhouFujian350118 China Fujian Key Lab for Automotive Electronics and Electric Drive Fujian University of Technology FuzhouFujian350118 China College of Humanities and Law Fujian Agriculture and Forestry University FuzhouFujian350002 China School of Mathematics and Information Science Baoji University of Arts and Sciences BaojiShaanxi721013 China

ISBN: (纸本)9783030037659

Visualization of ontology alignment and ontology entity is becoming a method to help Solve the problem of semantic heterogeneity. This paper first introduces the basic concepts of ontology and ontology matching, and then illustrates the importance of ontology visualization technique. The existing partial visualization tools are discussed respectively in two aspects of visualization of ontology alignment and visualization entity on ontology. Finally, four researching directions on ontology visualization technique are pointed out for the future work. © Springer Nature Switzerland AG 2019.

关键词： Ontology

来源：评论

学校读者我要写书评

暂无评论

A Dynamics Theory of Implicit Regularization in Deep Low-Rank Matrix Factorization

arXiv

引用

arXiv 2022年

作者： Cao, Jian Qian, Chen Huang, Yihui Chen, Dicheng Gao, Yuncheng Dong, Jiyang Guo, Di Qu, Xiaobo Department of Electronic Science Biomedical Intelligent Cloud R&D Center Fujian Provincial Key Laboratory of Plasma and Magnetic Resonance Xiamen University Xiamen361005 China School of Computer and Information Engineering Fujian Engineering Research Center for Medical Data Mining and Application Xiamen University of Technology Xiamen361024 China

Implicit regularization is an important way to interpret neural networks. Recent theory starts to explain implicit regularization with the model of deep matrix factorization (DMF) and analyze the trajectory of discrete gradient dynamics in the optimization process. These discrete gradient dynamics are relatively small but not infinitesimal, thus fitting well with the practical implementation of neural networks. Currently, discrete gradient dynamics analysis has been successfully applied to shallow networks but encounters the difficulty of complex computation for deep networks. In this work, we introduce another discrete gradient dynamics approach to explain implicit regularization, i.e. landscape analysis. It mainly focuses on gradient regions, such as saddle points and local minima. We theoretically establish the connection between saddle point escaping (SPE) stages and the matrix rank in DMF. We prove that, for a rank-R matrix reconstruction, DMF will converge to a second-order critical point after R stages of SPE. This conclusion is further experimentally verified on a low-rank matrix reconstruction problem. This work provides a new theory to analyze implicit regularization in deep learning. Copyright © 2022, 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

Studies and implementation of the trusted process mechanism in a secure operating system

引用

Qinghua Daxue Xuebao/Journal of Tsinghua University 2010年第SUPPL. 1期50卷 1547-1553页

作者： Zhang, Chongbin Li, Fengjuan Liang, Bin China Information Technology Security Evaluation Center Beijing 100085 China Key Laboratory of Data Engineering and Knowledge Engineering of Ministry of Education School of Information Renming University of China Beijing 100872 China

The trusted process mechanism is an important part of the operating system security mechanism, but there is no uniform definition of the trusted process and there are some limitations in understanding its fundamental properties. In this paper, the trustworthiness of a trusted process is discussed in terms of privilege, data-operated integrity, functional and logic correctness, self-integrity, availability and trusted interactions. A trusted process definition is then given which summarizes the fundamental properties of a trusted process as a theoretical foundation for researching and enforcing the trusted process mechanism. A prototype of a trusted process mechanism is implemented in a secure operating system based on Linux, with the trustworthiness of the trusted process assured in many ways to improve and extende the original trust mechanisms in Linux.

关键词： Linux

来源：评论

学校读者我要写书评

暂无评论

QGEval: Benchmarking Multi-dimensional Evaluation for Question Generation

arXiv

引用

arXiv 2024年

作者： Fu, Weiping Wei, Bifan Hu, Jianxiang Cai, Zhongmin Liu, Jun School of Computer Science and Technology Xi’an Jiaotong University Xi’an China School of Continuing Education Xi’an Jiaotong University Xi’an China MOE KLINNS Lab School of Automation Science and Engineering Xi’an Jiaotong University Xi’an China Shaanxi Province Key Laboratory of Big Data Knowledge Engineering Xi’an Jiaotong University Xi’an China

Automatically generated questions often suffer from problems such as unclear expression or factual inaccuracies, requiring a reliable and comprehensive evaluation of their quality. Human evaluation is widely used in the field of question generation (QG) and serves as the gold standard for automatic metrics. However, there is a lack of unified human evaluation criteria, which hampers consistent and reliable evaluations of both QG models and automatic metrics. To address this, we propose QGEval, a multi-dimensional Evaluation benchmark for Question Generation, which evaluates both generated questions and existing automatic metrics across 7 dimensions: fluency, clarity, conciseness, relevance, consistency, answerability, and answer consistency. We demonstrate the appropriateness of these dimensions by examining their correlations and distinctions. Through consistent evaluations of QG models and automatic metrics with QGEval, we find that 1) most QG models perform unsatisfactorily in terms of answerability and answer consistency, and 2) existing metrics fail to align well with human judgments when evaluating generated questions across the 7 dimensions. We expect this work to foster the development of both QG technologies and their evaluation. Copyright © 2024, The Authors. All rights reserved.

关键词： Benchmarking

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：