检索结果-内蒙古大学图书馆

Residual diverse ensemble for long-tailed multi-label text classification

Science China(information Sciences) 2024年第11期67卷 92-105页

作者： Jiangxin SHI Tong WEI Yufeng LI National Key Laboratory for Novel Software Technology Nanjing University School of Artificial Intelligence Nanjing University School of Computer Science and Engineering Southeast University Key Laboratory of Computer Network and Information Integration Southeast UniversityMinistry of Education

Long-tailed multi-label text classification aims to identify a subset of relevant labels from a large candidate label set, where the training datasets usually follow long-tailed label distributions. Many of the previous studies have treated head and tail labels equally, resulting in unsatisfactory performance for identifying tail labels. To address this issue, this paper proposes a novel learning method that combines arbitrary models with two steps. The first step is the “diverse ensemble” that encourages diverse predictions among multiple shallow classifiers, particularly on tail labels, and can improve the generalization of tail *** second is the “error correction” that takes advantage of accurate predictions on head labels by the base model and approximates its residual errors for tail labels. Thus, it enables the “diverse ensemble” to focus on optimizing the tail label performance. This overall procedure is called residual diverse ensemble(RDE). RDE is implemented via a single-hidden-layer perceptron and can be used for scaling up to hundreds of thousands of labels. We empirically show that RDE consistently improves many existing models with considerable performance gains on benchmark datasets, especially with respect to the propensity-scored evaluation ***, RDE converges in less than 30 training epochs without increasing the computational overhead.

关键词： multi-label learning extreme multi-label learning long-tailed distribution multi-label text classification ensemble learning

来源：评论

学校读者我要写书评

暂无评论

Design of a ML-based trust prediction model using intelligent TrustBoxes in challenged networks

引用

International Journal of Systems, Control and Communications 2024年第3期15卷 209-234页

作者： Barai, Smritikona Kundu, Anindita Bhaumik, Parama Department of Computer Science and Engineering Heritage Institute of Technology Kolkata India Department of Software Systems School of Computer Science and Engineering Vellore Institute of Technology Tamil Nadu India Department of Information Technology Jadavpur University Kolkata India

Challenged networks (CNs) contain resource-constrained nodes deployed in regions where human intervention is difficult. Opportunistic networks (OppNets) are CNs with no predefined source-to-destination paths. Due to their inherent properties, CNs and OppNets are highly susceptible to black-hole (BH) attacks, resulting in degraded packet-delivery ratio. In this work, an ML-based trust prediction model (MLTPM) is proposed to identify potential BH nodes in OppNets. MLTPM uses a novel function to calculate the total-trust-value (TTV) of each node. Intelligent TrustBoxes are introduced in the network to identify possible BH nodes, using TTV, along with five more node-behaviour features. TrustBoxes reduce the computational overhead of the resource-constrained nodes. Three simulated scenarios are compared - no detection, non-ML-based detection, and MLTPM, each using epidemic, prophet, and spray-and-wait routing protocols. MLTPM performs best with spray-and-wait, exhibiting about 25.21% and 80% mean improvement in delivery-ratio and dropped-message numbers respectively, compared to non-ML-based detection. An overall 12.62% improvement in delivery-ratio and 26.7% improvement in dropped messages is observed using MLTPM, compared to the above-mentioned scenarios. © 2024 Inderscience Enterprises Ltd.

关键词： HTTP

来源：评论

学校读者我要写书评

暂无评论

Boosting Adversarial Training with Learnable Distribution

引用

computers, Materials & Continua 2024年第3期78卷 3247-3265页

作者： Kai Chen Jinwei Wang James Msughter Adeke Guangjie Liu Yuewei Dai School of Electronics and Information Engineering Nanjing University of Information Science and TechnologyNanjing210044China Key Laboratory of Intelligent Support Technology for Complex Environments Ministry of EducationNanjing210044China School of Computer and Software Nanjing University of Information Science and TechnologyNanjing210044China Nanjing Center for Applied Mathematics Nanjing211135China

In recent years,various adversarial defense methods have been proposed to improve the robustness of deep neural *** training is one of the most potent methods to defend against adversarial ***,the difference in the feature space between natural and adversarial examples hinders the accuracy and robustness of the model in adversarial *** paper proposes a learnable distribution adversarial training method,aiming to construct the same distribution for training data utilizing the Gaussian mixture *** distribution centroid is built to classify samples and constrain the distribution of the sample *** natural and adversarial examples are pushed to the same distribution centroid to improve the accuracy and robustness of the *** proposed method generates adversarial examples to close the distribution gap between the natural and adversarial examples through an attack algorithm explicitly designed for adversarial *** algorithm gradually increases the accuracy and robustness of the model by scaling ***,the proposed method outputs the predicted labels and the distance between the sample and the distribution *** distribution characteristics of the samples can be utilized to detect adversarial cases that can potentially evade the model *** effectiveness of the proposed method is demonstrated through comprehensive experiments.

关键词： Adversarial training feature space learnable distribution distribution centroid

来源：评论

学校读者我要写书评

暂无评论

Improved Double Deep Q Network-Based Task Scheduling Algorithm in Edge Computing for Makespan Optimization

引用

Tsinghua Science and technology 2024年第3期29卷 806-817页

作者： Lei Zeng Qi Liu Shigen Shen Xiaodong Liu School of Computer Science Nanjing University of Information Science and TechnologyNanjing 210044China School of Software Nanjing University of Information Science and TechnologyNanjing 210044China School of Information Engineering Huzhou UniversityHuzhou 313000China School of Computing Edinburgh Napier UniversityEdinburghEH105DTUK

Edge computing nodes undertake an increasing number of tasks with the rise of business ***,how to efficiently allocate large-scale and dynamic workloads to edge computing resources has become a critical *** study proposes an edge task scheduling approach based on an improved Double Deep Q Network(DQN),which is adopted to separate the calculations of target Q values and the selection of the action in two networks.A new reward function is designed,and a control unit is added to the experience replay unit of the *** management of experience data are also modified to fully utilize its value and improve learning *** learning agents usually learn from an ignorant state,which is *** such,this study proposes a novel particle swarm optimization algorithm with an improved fitness function,which can generate optimal solutions for task *** optimized solutions are provided for the agent to pre-train network parameters to obtain a better cognition *** proposed algorithm is compared with six other methods in simulation *** show that the proposed algorithm outperforms other benchmark methods regarding makespan.

关键词： edge computing task scheduling reinforcement learning makespan Double Deep Q Network(DQN)

来源：评论

学校读者我要写书评

暂无评论

Binary Program Vulnerability Mining Based on Neural Network

引用

computers, Materials & Continua 2024年第2期78卷 1861-1879页

作者： Zhenhui Li Shuangping Xing Lin Yu Huiping Li Fan Zhou Guangqiang Yin Xikai Tang Zhiguo Wang School of Information and Software Engineering University of Electronic Science and Technology of ChinaChengdu611731China School of Electrical and Computer Engineering University ofWaterlooWaterlooN2L 3G1Canada

software security analysts typically only have access to the executable program and cannot directly access the source code of the *** poses significant challenges to security *** it is crucial to identify vulnerabilities in such non-source code programs,there exists a limited set of generalized tools due to the low versatility of current vulnerability mining ***,these tools suffer from some *** terms of targeted fuzzing,the path searching for target points is not streamlined enough,and the completely random testing leads to an excessively large search ***,when it comes to code similarity analysis,there are issues with incomplete code feature extraction,which may result in information *** this paper,we propose a cross-platform and cross-architecture approach to exploit vulnerabilities using neural network obfuscation *** leveraging the Angr framework,a deobfuscation technique is introduced,along with the adoption of a VEX-IR-based intermediate language conversion *** combination allows for the unified handling of binary programs across various architectures,compilers,and compilation ***,binary programs are processed to extract multi-level spatial features using a combination of a skip-gram model with self-attention mechanism and a bidirectional Long Short-Term Memory(LSTM)***,the graph embedding network is utilized to evaluate the similarity of program *** on these similarity scores,a target function is determined,and symbolic execution is applied to solve the target *** solved content serves as the initial seed for targeted *** binary program is processed by using the de-obfuscation technique and intermediate language transformation method,and then the similarity of program functions is evaluated by using a graph embedding network,and symbolic execution is performed based on these similarity *** approach facilitates

关键词： Vulnerability mining de-obfuscation neural network graph embedding network symbolic execution

来源：评论

学校读者我要写书评

暂无评论

An attention-based channel estimation network with residual mechanism for 5G OFDM systems 4

An attention-based channel estimation network with residual ...

引用

4th International Conference on Electronic Communication, computer Science and technology, ECCST 2024

作者： Jiang, Jinjing Huang, Weipeng Zhang, Xiao Yang, Kun Peng, Tong School of Information Engineering Zhejiang Ocean University China School of Computer Science and Software Engineering Shenzhen Institute of Information Technology China

Due to the significant increase in data transmission speed and gradual increase in Doppler frequency shift, channel estimation accuracy has become one of the most prioritized considerations in many cases. Specifically, we propose an attention-based channel estimation network with a residual mechanism, named TRCE-Net. In our design, the use of the multi-head attention mechanism at the input of the encoder is to obtain richer feature details. A linear series decoder module is implemented by leveraging multiple residual networks with skip connections. This novel design is capable of improving the channel estimation accuracy after deepening the network while effectively preventing degradation. To lessen the number of parameters that the network needs to learn, we introduce the bi-linear interpolation mechanism to replace the traditional up-sampling layer to mitigate both computational overhead and model complexity. According to the simulation results, our proposed TRCE-Net is superior to other state-of-the-art methods with regard to estimation accuracy and robustness to different maximum Doppler shifts, such as ResNet and InterpolateNet. © Published under licence by IOP Publishing Ltd.

关键词： Orthogonal frequency division multiplexing

来源：评论

学校读者我要写书评

暂无评论

Optimized Autonomous Computing with Trusted Resilient Data Management for IoT Emerging Networks

引用

IEEE Sensors Journal 2025年第11期25卷 20778-20785页

作者： Shafique, Ayesha Cheng, Benmao Zhao, Bo Alwakid, Ghadah Naif Haq, Muhammad Inaam Ul Laboratory of Intelligent IOT Technology and Applications in Universities China Jouf University College of Computer and Information Sciences Department of Computer Science Al Jowf Sakaka72341 Saudi Arabia Quanzhou University of Information Engineering School of Computer Science and Software Engineering Quanzhou362000 China

The innovative city network integrates numerous computational and physical components to develop real-time systems. These systems can capture sensor data and distribute it to end stations. Most solutions have been presented based on the centralized computing paradigm, which effectively and systematically increases data flow;however, distributed wireless technologies and heterogeneous network services continue to raise significant research problems. These challenges lower the optimization criteria and reflect communication structure around the network edges. In this research, we proposed a sustainable development for smart networks using efficient big data management with collaborative decisions for network devices. It differs from most existing work in the mentioned aspect. It applies computational lightweight intelligence for forwarding collected data using mobile collectors and reduces the congestion flow between devices on the limited bandwidth of wireless links. Moreover, the energy load is efficiently managed with edge-driven methods, and the incorporation of trusted devices ensures the integrity of the smart network. It also tackles potential communication threats smartly. Based on the experiments conducted in Network Simulator (NS-3), the proposed model enhances the efficacy of smart networks for performance metrics with reliability and effective management of resources in Internet of Things (IoT) network. © 2025 IEEE.

关键词： Mathematical models Computational modeling Intelligent sensors Data aggregation Internet of Things Security Routing Load modeling Silicon Data models

来源：评论

学校读者我要写书评

暂无评论

DIEONet:Domain-Invariant information Extraction and Optimization Network for Visual Place Recognition

引用

computers, Materials & Continua 2025年第3期82卷 5019-5033页

作者： Shaoqi Hou Zebang Qin Chenyu Wu Guangqiang Yin Xinzhong Wang Zhiguo Wang School of Computer Science and Technology Xinjiang UniversityUrumqi830046China School of Information and Software Engineering University of Electronic Science and Technology ofChinaChengdu611731China Institute of Public Security Kash Institute of Electronics and Information IndustryKashi844000China

Visual Place Recognition(VPR)technology aims to use visual information to judge the location of agents,which plays an irreplaceable role in tasks such as loop closure detection and *** is well known that previous VPR algorithms emphasize the extraction and integration of general image features,while ignoring the mining of salient features that play a key role in the discrimination of VPR *** this end,this paper proposes a Domain-invariant information Extraction and Optimization Network(DIEONet)for *** core of the algorithm is a newly designed Domain-invariant information Mining Module(DIMM)and a Multi-sample Joint Triplet Loss(MJT Loss).Specifically,DIMM incorporates the interdependence between different spatial regions of the feature map in the cascaded convolutional unit group,which enhances the model’s attention to the domain-invariant static object *** Loss introduces the“joint processing of multiple samples”mechanism into the original triplet loss,and adds a new distance constraint term for“positive and negative”samples,so that the model can avoid falling into local optimum during *** demonstrate the effectiveness of our algorithm by conducting extensive experiments on several authoritative *** particular,the proposed method achieves the best performance on the TokyoTM dataset with a Recall@1 metric of 92.89%.

关键词： Visual place recognition domain-invariant information mining module multi-sample joint triplet loss

来源：评论

学校读者我要写书评

暂无评论

A systematic mapping to investigate the application of machine learning techniques in requirement engineering activities

引用

CAAI Transactions on Intelligence technology 2024年第6期9卷 1412-1434页

作者： Shoaib Hassan Qianmu Li Khursheed Aurangzeb Affan Yasin Javed Ali Khan Muhammad Shahid Anwar School of Computer Science and Engineering Nanjing University of Science and TechnologyNanjingJiangsuChina Department of Computer Engineering College of Computer and Information SciencesKing Saud UniversityRiyadhSaudi Arabia School of Software Northwestern Polytechnical UniversityXianShaanxiChina Department of Computer Science School of PhysicsEngineering&Computer ScienceUniversity of HertfordshireHatfieldUK Department of AI and Software Gachon University Seongnam-siSeongnamSouth Korea

Over the past few years,the application and usage of Machine Learning(ML)techniques have increased exponentially due to continuously increasing the size of data and computing *** the popularity of ML techniques,only a few research studies have focused on the application of ML especially supervised learning techniques in Requirement engineering(RE)activities to solve the problems that occur in RE *** authors focus on the systematic mapping of past work to investigate those studies that focused on the application of supervised learning techniques in RE activities between the period of 2002–*** authors aim to investigate the research trends,main RE activities,ML algorithms,and data sources that were studied during this ***-five research studies were selected based on our exclusion and inclusion *** results show that the scientific community used 57 *** those algorithms,researchers mostly used the five following ML algorithms in RE activities:Decision Tree,Support Vector Machine,Naïve Bayes,K-nearest neighbour Classifier,and Random *** results show that researchers used these algorithms in eight major RE *** activities are requirements analysis,failure prediction,effort estimation,quality,traceability,business rules identification,content classification,and detection of problems in requirements written in natural *** selected research studies used 32 private and 41 public data *** most popular data sources that were detected in selected studies are the Metric Data Programme from NASA,Predictor Models in software engineering,and iTrust Electronic Health Care System.

关键词： data sources machine learning requirement engineering supervised learning algorithms

来源：评论

学校读者我要写书评

暂无评论

A dynamic and resource sharing virtual network mapping algorithm

引用

Digital Communications and Networks 2023年第5期9卷 1101-1112页

作者： Xiancui Xiao Xiangwei Zheng Ji Bian Cun Ji Xinchun Cui School of Information Science and Engineering Shandong Normal UniversityJinan250014China School of Information Engineering Shandong Management UniversityJinan250358China Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology Jinan250014China School of Computer Science and Technology Qufu Normal UniversityRizhao276800China

Network virtualization can effectively establish dedicated virtual networks to implement various network ***,the existing research works have some shortcomings,for example,although computing resource properties of individual nodes are considered,node storage properties and the network topology properties are usually ignored in Virtual Network(VN)modelling,which leads to the inaccurate measurement of node availability and *** addition,most static virtual network mapping methods allocate fixed resources to users during the entire life cycle,and the users’actual resource requirements vary with the workload,which results in resource allocation *** on the above analysis,in this paper,we propose a dynamic resource sharing virtual network mapping algorithm named NMA-PRS-VNE,first,we construct a new,more realistic network framework in which the properties of nodes include computing resources,storage resources and topology *** the node mapping process,three properties of the node are used to measure its mapping ***,we consider the resources of adjacent nodes and links instead of the traditional method of measuring the availability and priority of nodes by considering only the resource properties,so as to more accurately select the physical mapping nodes that meet the constraints and conditions and improve the success rate of subsequent link ***,we divide the resource requirements of Virtual Network Requests(VNRs)into basic subrequirements and variable sub-variable requirements to complete dynamic resource *** former represents monopolizing resource requirements by the VNRs,while the latter represents shared resources by many VNRs with the probability of occupying resources,where we keep a balance between resource sharing and collision among users by calculating the collision *** results show that the proposed NMAPRS-VNE can increase the average acceptance rate and network revenu

关键词： Network virtualization VNRs Network frameworks Dynamic resource allocation Resource sharing

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：