检索结果-内蒙古大学图书馆

arXiv 2022年

作者： He, Yin-Yin Zhang, Peizhen Wei, Xiu-Shen Zhang, Xiangyu Sun, Jian State Key Laboratory for Novel Software Technology Nanjing University China MEGVII Technology China School of Computer Science and Engineering Nanjing University of Science and Technology China

Long-tailed instance segmentation is a challenging task due to the extreme imbalance of training samples among classes. It causes severe biases of the head classes (with majority samples) against the tailed ones. This renders "how to appropriately define and alleviate the bias" one of the most important issues. Prior works mainly use label distribution or mean score information to indicate a coarse-grained bias. In this paper, we explore to excavate the confusion matrix, which carries the fine-grained misclassification details, to relieve the pairwise biases, generalizing the coarse one. To this end, we propose a novel Pairwise Class Balance (PCB) method, built upon a confusion matrix which is updated during training to accumulate the ongoing prediction preferences. PCB generates fightback soft labels for regularization during training. Besides, an iterative learning paradigm is developed to support a progressive and smooth regularization in such debiasing. PCB can be plugged and played to any existing method as a complement. Experimental results on LVIS demonstrate that our method achieves state-of-the-art performance without bells and whistles. Superior results across various architectures show the generalization ability. The code and trained models are available at https://***/megvii-research/PCB. Copyright © 2022, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Enhancing Federated Learning by One-Shot Transferring of Intermediate Features from Clients

Enhancing Federated Learning by One-Shot Transferring of Int...

引用

International Conference on Data Science and Advanced Analytics (DSAA)

作者： Youxingzhu Deng Yipeng Zhou Gang Liu Jessie Hui Wang Yu Shui College of Computer Science and Software Engineering Shenzhen University Shenzhen China School of Computing Faculty of Science and Engineering Macquarie University Sydney Australia Institute for Network Sciences and Cyberspace Tsinghua University Peking China School of Computer Science University of Technology Sydney Sydney Australia

Federated learning (FL) is an emerging paradigm using a parameter server (PS) to coordinate multiple decentralized clients for training a common model without exposing their raw data. Despite its amazing capability in preserving data privacy, FL confronts two significant challenges that have not been sufficiently addressed by existing works, which are: 1) heterogeneous data distributed on clients given that the PS cannot alter data locations. 2) limited computation resources of FL clients who may conduct model training with mobile devices. To tackle these challenges, we propose a novel federated one-shot transferring of intermediate features (FedOTF) algorithm. More specifically, FedOTF consists of two stages: feature extraction and model reconstruction. To overcome challenge 1, clients in stage 1 only collaboratively train a small model in a federated manner, with the objective to extract features. In stage 2, intermediate features (generated by the small model trained by stage 1) are transferred from clients to the PS so that a large model can be trained to overcome challenge 2. In addition, the design of FedOTF is robust, which can flexibly diminish the number of communication rounds of stage 1 when network capacity is limited, and reduce the amount of exposed features when privacy is concerned. To verify the superiority of FedOTF, we conduct comprehensive experiments with real datasets. The experiment results demonstrate that FedOTF can significantly improve the model utility of FL because a better model can be finally obtained by the PS without incurring heavy computational load on clients. Besides, we conduct robustness evaluation of FedOTF which can achieve stable performance when varying network capacity and privacy requirement.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Towards Planning Package Delivery Paths by Exploiting the Crowd of Private Cars

Towards Planning Package Delivery Paths by Exploiting the Cr...

引用

IEEE International Conference on High Performance Computing and Communications (HPCC)

作者： Hejing Nie Yiping Wen Tiancai Li Sipeng OuYang Muyang Tian School of Computer Science and Engineering Hunan University of Science and Technology China Hunan Key Laboratory for Service computing and Novel Software Technology Hunan University of Science & Technology

Crowdsourced delivery can provide cost-effective deliveries and it is becoming increasingly important in practice. This paper considers a package delivery problem which exploits relays of private cars to transport packages via their commuter routes. Specifically, we propose a method of planning package delivery paths based on the commuter routes of private cars. Using the commuter routes and time constraints as the reference, it can find optimal delivery paths and assign drivers of private cars to transport packages. The simulation experiment results show the validity and effectiveness of such method.

关键词： Product delivery Planning Automobiles Time factors Relays Task analysis

来源：评论

学校读者我要写书评

暂无评论

Automated information mining in hazardous chemical accident reporting: An improved deep learning approach

引用

Journal of Loss Prevention in the Process Industries 2025年 97卷

作者： Zhao, Kai Wan, Lining Lu, Xilei Zhao, Jun Chen, Fei He, Miao Gao, Jinhao Wang, Qibo Zhang, Linlin Zhang, Li School of Computer Science and Technology Xinjiang University Urumqi830017 China Xinjiang Uygur Autonomous Region Academy of Safety Science and Technology Urumqi830011 China Xinjiang Yihua Chemical Co. Changji Hui Autonomous Prefecture 831100 China School of Chemical Engineering and Technology Xinjiang University Urumqi830017 China School of Software Xinjiang University Urumqi830000 China

Mining and accumulating lessons learned from incident reports, accurately identifying and extracting accident knowledge, can help managers recognize patterns, analyze common attributes, and thus prevent the recurrence of similar incidents. Due to the limitations of unstructured or semi-structured text records, current hazardous chemical accident research heavily relies on expert evaluations, often resulting in inefficiency, lack of intelligence, and subjectivity. To overcome this limitation, we propose an accident analysis framework that combines deep learning-based Named Entity Recognition (NER) and topic mining algorithms. First, we employ a RoBERTa-BiLSTM-Attention-CRF method based on a pre-trained model to automatically recognize and extract information from accident reports according to predefined accident entities. Second, we construct a dictionary specific to the hazardous chemicals domain and apply the Latent Dirichlet Allocation (LDA) model to analyze the causes of accidents within the "CAUSE" entity, thereby deriving the hazardous themes implied in the accidents. Finally, we validate the model on a self-constructed hazardous chemical named entity recognition incident report dataset (HCNER). The research findings indicate that the methodology proposed in this study effectively addresses the challenges associated with dispersed textual information, specialized terminology, and data formatting in the digital processing of hazardous chemical incidents. This advancement propels the development within the domain of hazardous chemical safety management and serves as a viable approach for investigating the causes of such incidents. © 2025 Elsevier Ltd

关键词： Chemical hazards

来源：评论

学校读者我要写书评

暂无评论

Rare and Common Types of Retinal Disease Recognition Using Ensemble Deep Learning

Rare and Common Types of Retinal Disease Recognition Using E...

引用

International Conference on Electrical, computer and Communication engineering (ECCE)

作者： Jesika Debnath Al Shahriar Uddin Khondakar Pranta Arafath Bin Mohiuddin Amira Hossain Anamul Sakib Md. Moniruzzaman Shanto Shamim Forhad Abdullah Al-Sakib Rezaul Haque Shafiur Rahman Dept. of Computer Science Westcliff University Irvine CA USA Dept. of Computer Science Wright State University Dayton USA Department of Collage of Engineering and Technology Westcliff University CA USA Department of Business Administration International American University CA USA Dept. of Computer Science and Engineering East West University Dhaka Bangladesh Dept. of Electrical and Electronics Engineering American International University Bangladesh Dept. of Information Technology Westcliff University CA USA Dept. of Software Engineering Daffodil International University Dhaka Bangladesh

ISBN: (数字)9798350357509

ISBN: (纸本)9798350357516

Detecting retinal diseases is crucial for preventing vision loss, but it comes with challenges, such as data imbalance and accurately classifying rare conditions. This study introduces a new ensemble transfer learning model that combines EfficientNetB7, SE-ResNet, DenseNet-169, and NASNet to improve the classification of common and rare retinal diseases. The ensemble uses a probability averaging approach to maximise each architecture’s strengths. Essential preprocessing techniques such as Global Contrast Normalization (GCN) and Total Variation Denoising (TVD) were used to enhance image quality. The model was trained and evaluated on the Retinal Fundus Multi-Disease Image Dataset (RFMiD) 2.0, achieving 98.65% accuracy and an F1-score of 0.96% for common diseases and 99.10% accuracy with an F1-score of 0.97 for rare diseases, outperforming individual models significantly. The ensemble has been integrated into a web-based application that allows clinicians to upload retinal images for instant diagnostic feedback, enabling real-time analysis and visual insights. Despite its strengths, challenges such as high computational demands and reliance on high-resolution photos remain. Future work will improve computational efficiency through model compression techniques and extend the model to other imaging modalities, such as Optical Coherence Tomography (OCT). In conclusion, this study presents a robust and scalable solution for retinal disease classification, potentially improving early detection and providing real-time support in clinical settings, particularly in resource-limited environments.

关键词： Deep learning Visualization Accuracy Computational modeling Optical coherence tomography Transfer learning Imaging Retina Real-time systems Diseases

来源：评论

学校读者我要写书评

暂无评论

Enabling Communication-Efficient Federated Learning via Distributed Compressed Sensing

Enabling Communication-Efficient Federated Learning via Dist...

引用

IEEE Annual Joint Conference: INFOCOM, IEEE computer and Communications Societies

作者： Yixuan Guan Xuefeng Liu Tao Ren Jianwei Niu State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and Engineering Beihang University Beijing China Zhongguancun Laboratory Beijing China Laboratory for Internet Software Technologies Institute of Software Chinese Academy of Sciences Beijing China Zhengzhou University Research Institute of Industrial Technology School of Information Engineering Zhengzhou University Zhengzhou China

Federated learning (FL) trains a shared global model by periodically aggregating gradients from local devices. Communication overhead becomes a principal bottleneck in FL since participating devices usually suffer from limited bandwidth and unreliable connections in uplink transmission. To address this problem, the gradient compression methods based on compressed sensing (CS) theory have been put forward recently. However, most existing CS-based works compress gradients independently, ignoring the gradient correlations between participants or adjacent communication rounds, which constrains the achievement of higher compression rates. In view of the above observation, we propose a novel gradient compression scheme named FedDCS, guided by distributed compressed sensing (DCS) theory. Following the design philosophy of separate encoding and joint decoding in DCS, FedDCS compresses gradients for participants in each round separately while reconstructing them at the central server jointly via fully exploiting correlated gradients from the previous round, which are known as side information (SI). Benefiting from this design, reconstruction performance is significantly improved with fewer decoding errors also iterations under the identical compression rate, and the total uploading bits to achieve model convergence are considerably reduced. Theoretical analysis and extensive experiments conducted on MNIST and Fashion-MNIST both verify the effectiveness of our approach.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of a Following Vehicle Based on ROS and UWB

Design and Implementation of a Following Vehicle Based on RO...

引用

Automation, Control and Robots (ICACR), International Conference on

作者： Yun-Shuai Yu Chin-Wei Chen Department of Computer Science and Information Engineering National Formosa University Yunlin Taiwan Software Research & Development Center Hon Hai Technology Group (Foxconn) New Taipei City Taiwan

Autonomous vehicle technology has gained significant attention in various industries recently. This paper presents the design and implementation of a following vehicle that utilizes the Robot Operating System (ROS) and Ultra-Wideband (UWB) RF transceiver modules to achieve precise target localization, map the surrounding environment, and navigate accordingly. A mobile application has also been developed to track the following vehicle’s position.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Improving Tese Case Generation for Python Native Libraries Through Constraints on Input Data Structures

arXiv

引用

arXiv 2022年

作者： Zhang, Xin Ma, Xutong Yan, Jiwei Cui, Baoquan Yan, Jun Zhang, Jian State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences China Technology Center of Software Engineering Institute of Software Chinese Academy of Sciences China University of Chinese Academy of Sciences China

Modern Python projects execute computational functions using native libraries and give Python interfaces to boost execution speed;hence, testing these libraries becomes critical to the project's robustness. One challenge is that existing approaches use coverage to guide generation, but native libraries run as black boxes to Python code with no execution information. Another is that dynamic binary instrumentation reduces testing performance as it needs to monitor both native libraries and the Python virtual machine. To address these challenges, in this paper, we propose an automated test case generation approach that works at the Python code layer. Our insight is that many path conditions in native libraries are for processing input data structures through interacting with the VM. In our approach, we instrument the Python Interpreter to monitor the interactions between native libraries and VM, derive constraints on the structures, and then use the constraints to guide test case generation. We implement our approach in a tool named PyCing and apply it to six widely-used Python projects. The experimental results reveal that with the structure constraint guidance, PyCing can cover more execution paths than existing test cases and state-of-the-art tools. Also, with the checkers in the testing framework Pytest, PyCing can identify segmentation faults in 10 Python interfaces and memory leaks in 9. Our instrumentation strategy also has an acceptable influence on testing efficiency. © 2022, CC BY-NC-SA.

关键词： Python

来源：评论

学校读者我要写书评

暂无评论

Reliability of Multipath Networks with Optimization of the Location of Inter-Path Communication Nodes

Reliability of Multipath Networks with Optimization of the L...

引用

International Russian Smart Industry Conference (SmartIndustryCon)

作者： Vladimir A. Bogatyrev Anh Tu Le Evgenia A. Abramova Department Information Systems Security Saint-Petersburg State University of Aerospace Instrumentation Saint Petersburg Russia Faculty of Software Engineering and Computer Technology University ITMO Saint Peterburg Russia

The article explores the possibility of improving the reliability of a network with multipath routing. A feature of the proposed study is the analysis of the influence of the placement of switching nodes that switch path segments on the probability of network connectivity with servers of the same type and heterogeneous in functionality. The purpose of the article is to increase the reliability of the network, taking into account the influence of the placement of communication nodes that implement the switching of route segments during reconfiguration, on the connectivity of the network with servers that are homogeneous and heterogeneous in functionality. The research involves the construction of a structural reliability model that takes into account the possibility of failures of both communication nodes and the links between them. The reliability model is built taking into account the availability of request sources with full and non-full access connections to the set of network paths.

关键词： Industries Analytical models Switches Routing Reliability engineering Reliability Servers

来源：评论

学校读者我要写书评

暂无评论

Evaluation of Valued Tolerance Rough Set and Decision Rules Method for WiFi-Based Indoor Localization in Different Environments

Evaluation of Valued Tolerance Rough Set and Decision Rules ...

引用

Proceedings of the International Conference on Intelligent Systems and Network, ICISN 2023

作者： Duong-Bao, Ninh He, Jing Thi, Luong Nguyen Lee, Seon-Woo Nguyen-Huu, Khanh College of Computer Science and Electronic Engineering Hunan University Changsha410082 China Faculty of Mathematics and Informatics Dalat University Dalat66100 Viet Nam Faculty of Information Technology Dalat University Dalat66100 Viet Nam Division of Software Hallym University Chuncheon24252 Korea Republic of Department of Electronics and Telecommunications Dalat University Dalat66100 Viet Nam

ISBN: (纸本)9789819947249

Among various technologies being applied for indoor localization, WiFi has become a common source of information to determine the pedestrian’s position due to the widespread of WiFi access points in indoor environments. However, the fluctuation of the WiFi signals makes it difficult to achieve a good localization result. In this paper, to handle this problem, the valued tolerance rough set and decision rules method (VTRS-DR), which is firstly registered to WiFi fingerprinting-based localization, will be implemented and evaluated in big and complicated environments using two public datasets. The first one was conducted by a subject using a smartphone at a multi-floor library for several months. Furthermore, to evaluate the localization accuracy when WiFi data was collected from different pedestrians as well as different smartphones, a crowdsourced WiFi fingerprinting dataset was utilized. From the deep analyses of localization results, the VTRS-DR method shows high accuracy and high robustness when testing in different environments with a mean error of 3.84 m, which is 27.87% lower than other compared methods. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Smartphones

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：