检索结果-内蒙古大学图书馆

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Hong, Xin Gong, Yuan Sethu, Vidhyasaharan Dang, Ting School of Computing and Information Systems University of Melbourne Melbourne Australia CSAIL SLS Massachusetts Institute of Technology Boston United States School of Electrical Engineering and Telecommunications University of New South Wales Sydney Australia

ISBN: (纸本)9798350368741

Recent advancements in Large Language Models (LLMs) have demonstrated great success in many Natural Language Processing (NLP) tasks. In addition to their cognitive intelligence, exploring their capabilities in emotional intelligence is also crucial, as it enables more natural and empathetic conversational AI. Recent studies have shown LLMs' capability in recognizing emotions, but they often focus on single emotion labels and overlook the complex and ambiguous nature of human emotions. This study is the first to address this gap by exploring the potential of LLMs in recognizing ambiguous emotions, leveraging their strong generalization capabilities and in-context learning. We design zero-shot and few-shot prompting and incorporate past dialogue as context information for ambiguous emotion recognition. Experiments conducted using three datasets indicate significant potential for LLMs in recognizing ambiguous emotions, and highlight the substantial benefits of including context information. Furthermore, our findings indicate that LLMs demonstrate a high degree of effectiveness in recognizing less ambiguous emotions and exhibit potential for identifying more ambiguous emotions, paralleling human perceptual capabilities. © 2025 IEEE.

关键词： ambiguous emotion emotion recognition large language models multimodal prompt design

来源：评论

学校读者我要写书评

暂无评论

Performance and Accuracy Assessment of Detecting Network Intrusions with eSOM-Based Techniques

Performance and Accuracy Assessment of Detecting Network Int...

引用

2024 International Conference Automatics and Informatics, ICAI 2024

作者： Simeonidis, Dimitrios Petrov, Pavel Penchev, Georgi Petrova, Stefka Dimitrov, Georgi Petrivskyi, Volodymyr University of Economics - Varna Department of Informatics Varna Bulgaria College of Tourism Department of Tourism Varna Bulgaria University of Library Studies and Information Technologies Department of Information Technology Sofia Bulgaria Taras Shevchenko National University Department of Software Systems and Technologies Kyiv Ukraine

ISBN: (纸本)9798350353907

In this study are provided results about performance evaluation of Emergent Self-Organizing Map-based Intrusion Detection System in simulated wireless network environment. The experiment included detection of various network intrusions, including DoS, Probe, R2L, and U2R attacks. The results show that DoS (95%) and Probe (88%) attacks had the higher detection rates, while R2L (75%) and U2R (70%) attacks had the lower detection rates. The evaluated system demonstrated a low false positive rate of 5% and quick response times for overt attacks (2 seconds), though response times increased for covert attacks (4 seconds). During the experiment, the amount of used calculation resources was considerably high - the average CPU utilization was 70% and 80%, respectively. The results show that in order to effectively use this type of Intrusion Detection System for real company network security, there is need for further optimization. © 2024 IEEE.

关键词： Network intrusion

来源：评论

学校读者我要写书评

暂无评论

Imbalanced Graph Classification with Multi-scale Oversampling Graph Neural Networks

arXiv

引用

arXiv 2024年

作者： Ma, Rongrong Pang, Guansong Chen, Ling Faculty of Engineering and Information Technology University of Technology Sydney Sydney Australia School of Computing and Information Systems Singapore Management University Singapore

One main challenge in imbalanced graph classification is to learn expressive representations of the graphs in underrepresented (minority) classes. Existing generic imbalanced learning methods, such as oversampling and imbalanced learning loss functions, can be adopted for enabling graph representation learning models to cope with this challenge. However, these methods often directly operate on the graph representations, ignoring rich discriminative information within the graphs and their interactions. To tackle this issue, we introduce a novel multi-scale oversampling graph neural network (MOSGNN) that learns expressive minority graph representations based on intra- and inter-graph semantics resulting from oversampled graphs at multiple scales - subgraph, graph, and pairwise graphs. It achieves this by jointly optimizing subgraph-level, graph-level, and pairwise-graph learning tasks to learn the discriminative information embedded within and between the minority graphs. Extensive experiments on 16 imbalanced graph datasets show that MOSGNN i) significantly outperforms five state-of-the-art models, and ii) offers a generic framework, in which different advanced imbalanced learning loss functions can be easily plugged in and obtain significantly improved classification performance. Copyright © 2024, The Authors. All rights reserved.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Impersonated Human Speech Chatbot with Adaptive Frequency Spectrum 3

Impersonated Human Speech Chatbot with Adaptive Frequency Sp...

引用

3rd International Conference on Intelligent Communication and Computational Techniques, ICCT 2023

作者： Chettiar, Gautam Shukla, Amogh Nalwaya, Preet Sethi, Karman Prakash, Surya Vellore Institute of Technology Dept. of Communication Engineering Vellore India Department of Software Systems Vellore Institute of Technology Vellore India Indian Institute of Information Technology Allahabad Department of Electronics and Communication Engineering Prayagraj India

ISBN: (纸本)9781665453578

Recent trends in artificial intelligence and natural language processing models have led to the generation of highly efficient and versatile intelligent chatbot models., which have the potential to supplant human speech to the level of conversational proficiency. The proposed method creates a chatbot model that trains itself on open conversation datasets and aims to impersonate without compromising the emotional sentiments in the voice. These datasets extract from the applications such as WhatsApp., Telegram., Messenger., or any other chatting platform. Datasets convert to a machine-readable format., which is dynamically updated in real-time during the conversation., and then using speech conversion algorithms convert the reply into the desired individual's voice. The proposed model's conversational ability depends on the amount of conversation data., which gives the output in the person's voice frequency. By using an NLP-based chatbot trained on personalized data using KNN., and handling misses by pipelining the chatbot inputs to the GPT-2 model., the model can generate human-like replies even if there is data insufficiency. The natural replies are complemented with matching human voice and tone characteristics by using the vocoder model., which matches the spectral characteristics of the target voice onto the required voice. This opens a plethora of commercial and therapeutic applications that provide excellent insights into implementing natural communication models for humanoid and robotics innovations. © 2023 IEEE.

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

Energy-Saving VM in SDN-Driven Cloud Data Center with the Operational Vacation Policy 3

Energy-Saving VM in SDN-Driven Cloud Data Center with the Op...

引用

3rd International Conference on Innovative Mechanisms for Industry Applications, ICIMIA 2023

作者： Dash, Bibhuti Bhusan Patra, Sudhansu Shekhar Barik, Lalbihari Ghosh, Akash Pandey, Trilok Nath Gourisaria, Mahendra Kumar KIIT Deemed to Be University School of Computer Applications Bhubaneswar India King Abdul Aziz University Faculty of Computing and Information Technology in Rabigh Department of Information Systems Saudi Arabia Vellore Institute of Technology School of Computer Science and Engineering Chennai India

ISBN: (纸本)9798350343632

The increasing use of the Internet of Things (IoT), which has resulted in an exponential growth in network traffic, has serious implications on energy consumption and network performance. To reduce power consumption in cloud data centers, a server vacation-based energy-efficient traffic management solution is recommended. The primary idea is that using software-defined network (SDN) technologies, cloud system virtual machines (VMs) can take vacations. The goal of the energy-saving optimization task is to reduce the number of virtual machines (VMs) in use while maintaining service quality standards. This research study focuses on impatient client behaviour, several vacations, limited buffer and queuing system with many servers. Impatient customers may be kept in the cloud system by implementing certain customer retention strategies. The steady-state probability for the proposed SDN-assisted cloud system have been determined using recurrences. To determine the efficiency of the proposed system, several performance studies are conducted. © 2023 IEEE.

关键词： Cloud data center Energy optimization Performance measurement Queueing model software Defined Network Vacation policy

来源：评论

学校读者我要写书评

暂无评论

INCREASE: Inductive Graph Representation Learning for Spatio-Temporal Kriging 23

INCREASE: Inductive Graph Representation Learning for Spatio...

引用

32nd ACM World Wide Web Conference, WWW 2023

作者： Zheng, Chuanpan Fan, Xiaoliang Wang, Cheng Qi, Jianzhong Chen, Chaochao Chen, Longbiao Fujian Key Laboratory of Sensing and Computing for Smart Cities School of Informatics Xiamen University Xiamen China School of Computing and Information Systems University of Melbourne Melbourne Australia College of Computer Science and Technology Zhejiang University Hangzhou China

ISBN: (纸本)9781450394161

Spatio-temporal kriging is an important problem in web and social applications, such as Web or Internet of Things, where things (e.g., sensors) connected into a web often come with spatial and temporal properties. It aims to infer knowledge for (the things at) unobserved locations using the data from (the things at) observed locations during a given time period of interest. This problem essentially requires inductive learning. Once trained, the model should be able to perform kriging for different locations including newly given ones, without retraining. However, it is challenging to perform accurate kriging results because of the heterogeneous spatial relations and diverse temporal patterns. In this paper, we propose a novel inductive graph representation learning model for spatio-temporal kriging. We first encode heterogeneous spatial relations between the unobserved and observed locations by their spatial proximity, functional similarity, and transition probability. Based on each relation, we accurately aggregate the information of most correlated observed locations to produce inductive representations for the unobserved locations, by jointly modeling their similarities and differences. Then, we design relation-aware gated recurrent unit (GRU) networks to adaptively capture the temporal correlations in the generated sequence representations for each relation. Finally, we propose a multi-relation attention mechanism to dynamically fuse the complex spatio-temporal information at different time steps from multiple relations to compute the kriging output. Experimental results on three real-world datasets show that our proposed model outperforms state-of-the-art methods consistently, and the advantage is more significant when there are fewer observed locations. Our code is available at https://***/zhengchuanpan/INCREASE. © 2023 ACM.

关键词： Kriging

来源：评论

学校读者我要写书评

暂无评论

Offloading Decision Making in Mobile Edge computing: A Survey 2

Offloading Decision Making in Mobile Edge Computing: A Surve...

引用

2nd International Conference on Emerging Smart Technologies and Applications, eSmarTA 2022

作者： Abu-Taleb, Najlaa Abdulrahman Hasan Abdulrazzak, Fuad Zahary, Ammar Thabit Al-Mqdashi, Abdulmajid M. Thamar University Faculty of Computer Science and Information Systems Thamar Yemen University of Science and Technology Faculty of Computing and IT Sana'a Yemen

ISBN: (数字)9781665461818

ISBN: (纸本)9781665461818

Mobile Edge computing (MEC) is considered as a promising solution to augment the computational capabilities of mobile devices. Because of frequently changing locations, mobile users often see dynamically changing network conditions which have a great impact on the perceived application performance. Mobile users can offload computation-intensive tasks to edge clouds collocated with base stations. Task offloading in MEC is a solution to handle such computational-intensive applications. However, making high-quality offloading decisions at runtime is difficult in mobile environments. This paper presents a survey for offloading decisions in MEC. It shows the most critical issues and challenges that affect offloading decisions. In addition, the survey discusses the existing offloading decision-making approaches and models which have been categorized according to their schemes' optimization. Furthermore, a literature review of offloading decisions for various studies is investigated and analyzed from the perspectives of objectives, infrastructures, and optimization approach. Finally, the survey highlights the open research challenges and trends in the field of offloading decisions in MEC. © 2022 IEEE.

关键词： Mobile edge computing

来源：评论

学校读者我要写书评

暂无评论

An Aircraft Identification System Using Convolution Neural Networks 12

An Aircraft Identification System Using Convolution Neural N...

引用

12th IEEE International Conference on Intelligent Data Acquisition and Advanced computing systems: technology and Applications, IDAACS 2023

作者： Turchenko, Viktor Turchenko, Iryna Vasylkiv, Nadiia West Ukrainian National University Department for Information Computer Systems and Control Ternopil46000 Ukraine McMaster University Department of Computing and Software HamiltonL8S 4L8 Canada Research Institute for Intelligent Computer Systems West Ukrainian National University Ternopil Ukraine

ISBN: (纸本)9798350358056

The task of aircraft identification using their registration number is important as visual identification remains the only method of their identification in case of air traffic control systems or radar failure. In this paper, an analysis of known solutions was carried out and an aircraft identification system was developed using a pre-trained YOLO (You Only Look Once) deep convolutional neural network and Optical Character Recognition (OCR) technology. The system consists of three modules: the M1 module is designed for the detection of region of interest-tail, the M2 module is designed for the detection of region of interest-registration number, the M3 module is designed for the classification of the registration number. 2309 aircraft images were used for training, 426 for validation and 408 for testing from the Fine-Grained Visual Classification of Aircraft (FGVC-Aircraft) dataset. Our results show that the mean average precision for the M1 and M2 YOLOv7 models on the test set were 0.92 and 0.85 respectively, and for the module M3 OCR technology the classification error was 0.2. © 2023 IEEE.

关键词： aircraft convolutional neural networks deep learning detection transfer learning YOLOv7

来源：评论

学校读者我要写书评

暂无评论

Cluster Availability Model with information Update Following Node Recovery

Cluster Availability Model with Information Update Following...

引用

2023 Wave Electronics and its Application in information and Telecommunication systems, WECONF 2023

作者： Bogatyrev, V.A. Bogatyrev, S.V. Sivov, V.V. Saint-Petersburg State University of Aerospace Instrumentation Department Information Systems Security Saint Petersburg Russia ITMO University Faculty of Software Engineering and Computer Technology Saint-Petersburg Russia

ISBN: (纸本)9798350348293

A model for evaluating the availability of a fault-tolerant cluster with a constraint on the maximum query service time is proposed. Node recovery in the cluster involves replacing the entire computational node with the latest backup results before the failure. information update after the backup is achieved through the replication of this information from the working nodes of the cluster. The purpose of the article is to assess the reliability of the cluster and its readiness to timely service requests critical to the waiting time. Restoring up-to-date data in a cluster node after previously entering the data saved during backup into it requires the use of resources of the node in which the data is being restored and the node containing up-to-date data. The specified pair of nodes is excluded from the request servicing process, which reduces the likelihood of timely maintenance. information recovery in case of failures of all nodes containing replicas of up-to-date data requires more time compared to replication from the memory of a functioning node. © 2023 IEEE.

关键词： Fault tolerance

来源：评论

学校读者我要写书评

暂无评论

IDS-INT:Intrusion detection system using transformer-based transfer learning for imbalanced network traffic

引用

Digital Communications and Networks 2024年第1期10卷 190-204页

作者： Farhan Ullah Shamsher Ullah Gautam Srivastava Jerry Chun-Wei Lin School of Software Northwestern Polytechnical UniversityXian710072ShaanxiChina Knowledge Unit of Systems and Technology University of Management and TechnologySialkot 51040Pakistan Department of Math and Computer Science Brandon UniversityBrandonMB R7A 6A9Canada Department of Computer Science and Math Lebanese American UniversityBeirut1102Lebanon Western Norway University of Applied Sciences BergenNorway Research Centre for Interneural Computing China Medical UniversityTaichungTaiwanChina

A network intrusion detection system is critical for cyber security against llegitimate *** terms of feature perspectives,network traffic may include a variety of elements such as attack reference,attack type,a subcategory of attack,host information,malicious scripts,*** terms of network perspectives,network traffic may contain an imbalanced number of harmful attacks when compared to normal *** is challenging to identify a specific attack due to complex features and data imbalance *** address these issues,this paper proposes an Intrusion Detection System using transformer-based transfer learning for Imbalanced Network Traffic(IDS-INT).IDS-INT uses transformer-based transfer learning to learn feature interactions in both network feature representation and imbalanced ***,detailed information about each type of attack is gathered from network interaction descriptions,which include network nodes,attack type,reference,host information,***,the transformer-based transfer learning approach is developed to learn detailed feature representation using their semantic ***,the Synthetic Minority Oversampling Technique(SMOTE)is implemented to balance abnormal traffic and detect minority ***,the Convolution Neural Network(CNN)model is designed to extract deep features from the balanced network ***,the hybrid approach of the CNN-Long Short-Term Memory(CNN-LSTM)model is developed to detect different types of attacks from the deep *** experiments are conducted to test the proposed approach using three standard datasets,i.e.,UNsWNB15,CIC-IDS2017,and *** explainable AI approach is implemented to interpret the proposed method and develop a trustable model.

关键词： Network intrusion detection Transfer learning Features extraction Imbalance data Explainable AI Cybersecurity

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：