Pub Date : 2023-07-01DOI: 10.1109/CSCloud-EdgeCom58631.2023.00048
Benshan Mei, Lin Chen, Shao-Jie Sun, Pan-Yu Chen, Wei-Liang Huang
With the problems of handling single-feature and overlooking user preferences in the recommendation algorithms, this paper proposes a Recurrent Neural NetWork-based Multi-feature Hybrid Recommendation Model (RN-MHRM). Firstly, features are extracted from user-item interaction data using the Latent Factor Model (LFM), and an improved Recurrent Neural NetWork (RNN) is used to replace the linear inner product of LFM vectors With non-linearity, Which aims at learning richer features that capture user's short-term interests. Secondly, to avoid single-feature, item information is introduced and the BERT model is used for extracting multi-features. Thirdly, both short-term and long-term interests are considered, and the user's long-term interests are trained by a FeedforWard Neural NetWork (FNN), Which greatly improves the recommendation performance. Experiments designed on multiple real datasets have shown that RN-MHRM effectively improves recommendation performance compared to the baseline model.
{"title":"A New Multi-Feature Recommendation Model Based on Recurrent Neural Network","authors":"Benshan Mei, Lin Chen, Shao-Jie Sun, Pan-Yu Chen, Wei-Liang Huang","doi":"10.1109/CSCloud-EdgeCom58631.2023.00048","DOIUrl":"https://doi.org/10.1109/CSCloud-EdgeCom58631.2023.00048","url":null,"abstract":"With the problems of handling single-feature and overlooking user preferences in the recommendation algorithms, this paper proposes a Recurrent Neural NetWork-based Multi-feature Hybrid Recommendation Model (RN-MHRM). Firstly, features are extracted from user-item interaction data using the Latent Factor Model (LFM), and an improved Recurrent Neural NetWork (RNN) is used to replace the linear inner product of LFM vectors With non-linearity, Which aims at learning richer features that capture user's short-term interests. Secondly, to avoid single-feature, item information is introduced and the BERT model is used for extracting multi-features. Thirdly, both short-term and long-term interests are considered, and the user's long-term interests are trained by a FeedforWard Neural NetWork (FNN), Which greatly improves the recommendation performance. Experiments designed on multiple real datasets have shown that RN-MHRM effectively improves recommendation performance compared to the baseline model.","PeriodicalId":56007,"journal":{"name":"Journal of Cloud Computing-Advances Systems and Applications","volume":"52 1","pages":"235-240"},"PeriodicalIF":4.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90661901","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1109/CSCloud-EdgeCom58631.2023.00028
Zhu Kai, Bao-kang Zhao, Qin Xin
Flying Ad-hoc network (FANETs), a new kind of mobile AD hoc network, uses the aircraft as the air wireless communication node to construct a network and achieve effective communication at the network layer as a result of the ongoing expansion of communication networks into many industries. However, the high dynamics, limited node energy, and low network density of the FANETs present significant challenges to the FANETs routing protocol, which requires urgent attention in terms of its design. In order to address the issues associated with the routing metric of the FANETs routing protocol, such as the underutilization of link information and the inadequate consideration of the motion of UAVs, which result in high network packet loss rates, unstable routing, and long route reconvergence times, this paper proposes the Predictable Track-based Routing Protocol (PTP). PTP divides the routing protocol into three stages: route establishment, data transmission, and route maintenance. PTP creatively suggests Node Stability Factors (NSF) for UAV in the route establishment stage and develops a Q-learning algorithm to maintain it in accordance with the status of the nodes and nearby nodes. NSF combined the link quality calculated by HELLO messages to generate a link quality based on Q-learning(LQQ) metric. After that, the optimal path from the source node to the destination node is calculated based on this metric. In the data transfer phase, data is transmitted through the routes calculated in the route establishment phase. Create a new HELLO message during the routing maintenance phase to learn the location of the neighbor node and the two-hop neighbor node, predict the node’s future location using the Kalman filter algorithm based on the node’s past location, and react quickly when the link is about to change. Compared to conventional approaches, experiments demonstrate that PTP may successfully raise the successful data delivery rate by 10% to 30%, increase the average route survival time by up to 60%, and cut the average route reconvergence time by up to 80%.
飞行自组网(Flying AD -hoc network, fanet)是一种新型的移动自组网，它利用飞机作为空中无线通信节点来构建网络，实现网络层的有效通信，是通信网络不断向多个行业扩展的结果。然而，FANETs的高动态性、有限的节点能量和低网络密度对FANETs路由协议提出了重大挑战，需要在其设计方面引起迫切的重视。针对FANETs路由协议的路由度量存在链路信息利用不足、无人机运动考虑不足等问题，导致网络丢包率高、路由不稳定、路由重新收敛时间长等问题，提出了基于可预测轨迹的路由协议(PTP)。PTP将路由协议分为路由建立、数据传输和路由维护三个阶段。PTP创造性地提出了无人机在航路建立阶段的节点稳定因子(NSF)，并根据节点和附近节点的状态，开发了q -学习算法来维持NSF。NSF结合HELLO消息计算出的链路质量，生成基于Q-learning(LQQ)度量的链路质量。然后，根据该度量计算出从源节点到目的节点的最优路径。在数据传输阶段，数据通过路由建立阶段计算出的路由进行传输。在路由维护阶段创建新的HELLO消息，学习邻居节点和两跳邻居节点的位置，根据节点过去的位置使用卡尔曼滤波算法预测节点的未来位置，并在链路即将发生变化时快速反应。实验表明，与传统方法相比，PTP可以成功地将数据成功投递率提高10% ~ 30%，将平均路由生存时间提高60%，将平均路由重新收敛时间降低80%。
{"title":"Predictable Track-based Routing in Flying Ad hoc Networks","authors":"Zhu Kai, Bao-kang Zhao, Qin Xin","doi":"10.1109/CSCloud-EdgeCom58631.2023.00028","DOIUrl":"https://doi.org/10.1109/CSCloud-EdgeCom58631.2023.00028","url":null,"abstract":"Flying Ad-hoc network (FANETs), a new kind of mobile AD hoc network, uses the aircraft as the air wireless communication node to construct a network and achieve effective communication at the network layer as a result of the ongoing expansion of communication networks into many industries. However, the high dynamics, limited node energy, and low network density of the FANETs present significant challenges to the FANETs routing protocol, which requires urgent attention in terms of its design. In order to address the issues associated with the routing metric of the FANETs routing protocol, such as the underutilization of link information and the inadequate consideration of the motion of UAVs, which result in high network packet loss rates, unstable routing, and long route reconvergence times, this paper proposes the Predictable Track-based Routing Protocol (PTP). PTP divides the routing protocol into three stages: route establishment, data transmission, and route maintenance. PTP creatively suggests Node Stability Factors (NSF) for UAV in the route establishment stage and develops a Q-learning algorithm to maintain it in accordance with the status of the nodes and nearby nodes. NSF combined the link quality calculated by HELLO messages to generate a link quality based on Q-learning(LQQ) metric. After that, the optimal path from the source node to the destination node is calculated based on this metric. In the data transfer phase, data is transmitted through the routes calculated in the route establishment phase. Create a new HELLO message during the routing maintenance phase to learn the location of the neighbor node and the two-hop neighbor node, predict the node’s future location using the Kalman filter algorithm based on the node’s past location, and react quickly when the link is about to change. Compared to conventional approaches, experiments demonstrate that PTP may successfully raise the successful data delivery rate by 10% to 30%, increase the average route survival time by up to 60%, and cut the average route reconvergence time by up to 80%.","PeriodicalId":56007,"journal":{"name":"Journal of Cloud Computing-Advances Systems and Applications","volume":"45 1","pages":"114-119"},"PeriodicalIF":4.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82078062","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
With the widespread use of the Ride-on Demand (RoD) system, many privacy issues have been exposed, and there is growing concern about whether private information will be leaked. For this problem, our previous work addressed the issue of the user’s initial and final location leakage and provided a strong utility guarantee in the RoD system. Further, the trajectory information is also important in the RoD system, it could contain a lot of private information about the user, such as health or identity, so it’s important to publish a distorted and productive trajectory. For this purpose, in this paper, we provide our trajectory protection method with pricing awareness based on previous work; the method uses supply and demand density function to guide the division of a discrete spatial grid, then uses the Markov chain to generate distorted trajectories on the grid to ensure trajectory continuity, and makes corresponding defenses against several attacks, such as Bayesian. The experiment results on real-world datasets prove the validity and robustness of the method.
{"title":"Trajectory Privacy Protection with Pricing Awareness on Ride-on-Demand System","authors":"Sihui Jia, Saiqin Long, Z. Zheng, Qingyong Deng, Ping Wang, Shujuan Tian","doi":"10.1109/CSCloud-EdgeCom58631.2023.00016","DOIUrl":"https://doi.org/10.1109/CSCloud-EdgeCom58631.2023.00016","url":null,"abstract":"With the widespread use of the Ride-on Demand (RoD) system, many privacy issues have been exposed, and there is growing concern about whether private information will be leaked. For this problem, our previous work addressed the issue of the user’s initial and final location leakage and provided a strong utility guarantee in the RoD system. Further, the trajectory information is also important in the RoD system, it could contain a lot of private information about the user, such as health or identity, so it’s important to publish a distorted and productive trajectory. For this purpose, in this paper, we provide our trajectory protection method with pricing awareness based on previous work; the method uses supply and demand density function to guide the division of a discrete spatial grid, then uses the Markov chain to generate distorted trajectories on the grid to ensure trajectory continuity, and makes corresponding defenses against several attacks, such as Bayesian. The experiment results on real-world datasets prove the validity and robustness of the method.","PeriodicalId":56007,"journal":{"name":"Journal of Cloud Computing-Advances Systems and Applications","volume":"60 1","pages":"37-45"},"PeriodicalIF":4.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85627610","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1109/CSCloud-EdgeCom58631.2023.00063
Peiyuan Guan, Yushuai Li, Amirhosein Taherkordi
Handover is a crucial issue in ensuring the continuity of edge services in edge computing (EC) systems. Failure to handle hand-off properly may result in delays, data loss, or service interruption during service switching. Therefore, optimizing the hand-off process to ensure service continuity and satisfactory user experience is a significant challenge in the design of edge computing systems. In this paper, we propose a resource reservation algorithm that reserves a portion of computing resources in each base station to meet quality of service (QoS) requirements during service switching. We use an LSTM model to predict the number of new and existing users at a future time point to provide decision guidance for the resource reservation algorithm. Extensive simulation experiments demonstrate that the proposed algorithm outperforms the benchmark algorithms in a variety of environmental conditions.
{"title":"A Prediction Based Resource Reservation Algorithm for Service Handover in Edge Computing","authors":"Peiyuan Guan, Yushuai Li, Amirhosein Taherkordi","doi":"10.1109/CSCloud-EdgeCom58631.2023.00063","DOIUrl":"https://doi.org/10.1109/CSCloud-EdgeCom58631.2023.00063","url":null,"abstract":"Handover is a crucial issue in ensuring the continuity of edge services in edge computing (EC) systems. Failure to handle hand-off properly may result in delays, data loss, or service interruption during service switching. Therefore, optimizing the hand-off process to ensure service continuity and satisfactory user experience is a significant challenge in the design of edge computing systems. In this paper, we propose a resource reservation algorithm that reserves a portion of computing resources in each base station to meet quality of service (QoS) requirements during service switching. We use an LSTM model to predict the number of new and existing users at a future time point to provide decision guidance for the resource reservation algorithm. Extensive simulation experiments demonstrate that the proposed algorithm outperforms the benchmark algorithms in a variety of environmental conditions.","PeriodicalId":56007,"journal":{"name":"Journal of Cloud Computing-Advances Systems and Applications","volume":"99 1","pages":"330-335"},"PeriodicalIF":4.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73451820","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1109/CSCloud-EdgeCom58631.2023.00053
M. K. B. Shuhan, Tariqul Islam, Enam A. Shuvo, Faisal Haque Bappy, Kamrul Hasan, Carlos Caicedo
Over the past two decades, the popularity of messaging systems has increased both in enterprise and consumer level. Many of these systems used secure protocols like end-to-end encryption to ensure strong security features such as “future secrecy” for one-to-one communication. However, the majority of them rely on centralized servers owned by big IT companies, which allows them to use their users’ personal data. Also it allows the government to track and regulate their citizens’ activities, which poses significant threats to “digital freedom”. Also, these systems have failed to achieve security attributes like confidentiality, integrity, privacy, and future secrecy for group communications. In this paper, we present a novel blockchain-based secure messaging system named Quarks that overcomes the security pitfalls of the existing systems and eliminates the centralized control. We have analyzed our design of the system with security models and definitions from existing literature to demonstrate the system’s reliability and usability. We have developed a Proof of Concept (PoC) of the Quarks system leveraging Distributed Ledger Technology (DLT), and conducted load testing on that. We noticed that our PoC system achieves all the desired attributes that are prevalent in a traditional centralized messaging scheme despite the limited capacity of the development and testing environment. Therefore, this assures us the applicability of such systems in near future if scaled up properly.
{"title":"Quarks: A Secure and Decentralized Blockchain-Based Messaging Network","authors":"M. K. B. Shuhan, Tariqul Islam, Enam A. Shuvo, Faisal Haque Bappy, Kamrul Hasan, Carlos Caicedo","doi":"10.1109/CSCloud-EdgeCom58631.2023.00053","DOIUrl":"https://doi.org/10.1109/CSCloud-EdgeCom58631.2023.00053","url":null,"abstract":"Over the past two decades, the popularity of messaging systems has increased both in enterprise and consumer level. Many of these systems used secure protocols like end-to-end encryption to ensure strong security features such as “future secrecy” for one-to-one communication. However, the majority of them rely on centralized servers owned by big IT companies, which allows them to use their users’ personal data. Also it allows the government to track and regulate their citizens’ activities, which poses significant threats to “digital freedom”. Also, these systems have failed to achieve security attributes like confidentiality, integrity, privacy, and future secrecy for group communications. In this paper, we present a novel blockchain-based secure messaging system named Quarks that overcomes the security pitfalls of the existing systems and eliminates the centralized control. We have analyzed our design of the system with security models and definitions from existing literature to demonstrate the system’s reliability and usability. We have developed a Proof of Concept (PoC) of the Quarks system leveraging Distributed Ledger Technology (DLT), and conducted load testing on that. We noticed that our PoC system achieves all the desired attributes that are prevalent in a traditional centralized messaging scheme despite the limited capacity of the development and testing environment. Therefore, this assures us the applicability of such systems in near future if scaled up properly.","PeriodicalId":56007,"journal":{"name":"Journal of Cloud Computing-Advances Systems and Applications","volume":"2 1","pages":"268-274"},"PeriodicalIF":4.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87916950","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Synthetic aperture radar (SAR) satellites can provide microwave remote sensing images that are not limited by weather and light, so they are widely used in the field of ocean monitoring. The current SAR ship detection method based on deep learning (DL) is difficult to more effectively fuse complex features, which leads to low detection accuracy of target ships and even missed or false detections. In order to solve this problem, this paper proposes an improved YOLOX-based SAR image ship detection method, called NAS-YOLOX. Based on the YOLOX algorithm, path aggregation feature pyramid network (PAFPN) is replaced by a neural architecture search - feature fusion network (NAS-FPN) to enhance the cross-scale fusion ability of the proposed model. And a dilated convolution feature enhancement module (DFEM) is also designed and embedded into the backbone network to boost the network receptive field and the ability to extract target information. Furthermore, a multi-scale channel-spatial attention (MCSA) is proposed to improve the attention to key areas of the ship. The experimental results on the HRSID public data set show that the AP0.5 of NAS-YOLOX is 6.3% higher than that of the YOLOX model. Compared with other ten mainstream target detection algorithms, NAS-YOLOX has also achieved excellent detection result.
{"title":"NAS-YOLOX: ship detection based on improved YOLOX for SAR imagery","authors":"Hao Wang, Dezhi Han, Zhongdai Wu, Junxiang Wang, Yuan Fan, Yachao Zhou","doi":"10.1109/CSCloud-EdgeCom58631.2023.00030","DOIUrl":"https://doi.org/10.1109/CSCloud-EdgeCom58631.2023.00030","url":null,"abstract":"Synthetic aperture radar (SAR) satellites can provide microwave remote sensing images that are not limited by weather and light, so they are widely used in the field of ocean monitoring. The current SAR ship detection method based on deep learning (DL) is difficult to more effectively fuse complex features, which leads to low detection accuracy of target ships and even missed or false detections. In order to solve this problem, this paper proposes an improved YOLOX-based SAR image ship detection method, called NAS-YOLOX. Based on the YOLOX algorithm, path aggregation feature pyramid network (PAFPN) is replaced by a neural architecture search - feature fusion network (NAS-FPN) to enhance the cross-scale fusion ability of the proposed model. And a dilated convolution feature enhancement module (DFEM) is also designed and embedded into the backbone network to boost the network receptive field and the ability to extract target information. Furthermore, a multi-scale channel-spatial attention (MCSA) is proposed to improve the attention to key areas of the ship. The experimental results on the HRSID public data set show that the AP0.5 of NAS-YOLOX is 6.3% higher than that of the YOLOX model. Compared with other ten mainstream target detection algorithms, NAS-YOLOX has also achieved excellent detection result.","PeriodicalId":56007,"journal":{"name":"Journal of Cloud Computing-Advances Systems and Applications","volume":"10 1","pages":"126-131"},"PeriodicalIF":4.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86147868","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1109/CSCloud-EdgeCom58631.2023.00018
Ruoli Zhao, Yong Xie, Lijun Zhang, Haiyan Cao, Ping Liu
With the continuous development of Tibetan medicine, using machine learning technology to enhance the value of Tibetan medical data has become very important. However, the concerns of Tibetan medical institutions about data leakage have hindered the sharing of Tibetan medical data. Therefore, in this paper, we propose a privacy preserving computation framework based on dual servers. Our framework can securely store Tibetan medical data on cloud servers. The secure computation (such as machine learning training or machine learning prediction) is performed by cloud servers without compromising data. On the premise of ensuring data security, we combine the multi-key homomorphic encryption and secret sharing to design some secure building blocks. Through the security analysis and performance evaluation, our proposed scheme is efficient and practical.
{"title":"An Efficient and Privacy preserving Computation Framework for Tibetan medicine","authors":"Ruoli Zhao, Yong Xie, Lijun Zhang, Haiyan Cao, Ping Liu","doi":"10.1109/CSCloud-EdgeCom58631.2023.00018","DOIUrl":"https://doi.org/10.1109/CSCloud-EdgeCom58631.2023.00018","url":null,"abstract":"With the continuous development of Tibetan medicine, using machine learning technology to enhance the value of Tibetan medical data has become very important. However, the concerns of Tibetan medical institutions about data leakage have hindered the sharing of Tibetan medical data. Therefore, in this paper, we propose a privacy preserving computation framework based on dual servers. Our framework can securely store Tibetan medical data on cloud servers. The secure computation (such as machine learning training or machine learning prediction) is performed by cloud servers without compromising data. On the premise of ensuring data security, we combine the multi-key homomorphic encryption and secret sharing to design some secure building blocks. Through the security analysis and performance evaluation, our proposed scheme is efficient and practical.","PeriodicalId":56007,"journal":{"name":"Journal of Cloud Computing-Advances Systems and Applications","volume":"6 1","pages":"53-58"},"PeriodicalIF":4.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74628181","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1109/cscloud-edgecom58631.2023.00007
{"title":"Committee Members - CSCloud 2023","authors":"","doi":"10.1109/cscloud-edgecom58631.2023.00007","DOIUrl":"https://doi.org/10.1109/cscloud-edgecom58631.2023.00007","url":null,"abstract":"","PeriodicalId":56007,"journal":{"name":"Journal of Cloud Computing-Advances Systems and Applications","volume":"9 1","pages":""},"PeriodicalIF":4.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74565115","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1109/CSCloud-EdgeCom58631.2023.00054
Yufeng Xiao, Yuqin Bo, Zhiling Zheng
The application of semi-supervised learning in the field of speech emotion recognition can alleviate the problem of dependence on the amount of labeled data. Its core assumption is that the labeled and unlabeled data have the same feature representations, which determine the performance of the models. Therefore, this paper proposes a novel speech emotion recognition model based on semi-supervised adversarial variational autoencoding (SSAVAE), which reaps the advantages of generative adversarial network(GAN) and variational autoencoding(VAE) to learn the distribution of input data in the feature space. On the one hand, it can overcome the disturbance of the input data through learning the distribution of the input data in the feature space. On the other hand, it can approximate arbitrary feature distribution by introducing GAN. Concretely, SSAVAE considers the unlabeled data as the problem of the lack of emotional label attributes. The labeled data share the category information with the unlabeled data in the feature space. Several experiments are conducted on the FAU Aibo dataset to evaluate the effectiveness of the algorithm. The results show that the proposed method is superior to other benchmark algorithms and has strong feature learning capabilities.
{"title":"Speech Emotion Recognition based on Semi-Supervised Adversarial Variational Autoencoder","authors":"Yufeng Xiao, Yuqin Bo, Zhiling Zheng","doi":"10.1109/CSCloud-EdgeCom58631.2023.00054","DOIUrl":"https://doi.org/10.1109/CSCloud-EdgeCom58631.2023.00054","url":null,"abstract":"The application of semi-supervised learning in the field of speech emotion recognition can alleviate the problem of dependence on the amount of labeled data. Its core assumption is that the labeled and unlabeled data have the same feature representations, which determine the performance of the models. Therefore, this paper proposes a novel speech emotion recognition model based on semi-supervised adversarial variational autoencoding (SSAVAE), which reaps the advantages of generative adversarial network(GAN) and variational autoencoding(VAE) to learn the distribution of input data in the feature space. On the one hand, it can overcome the disturbance of the input data through learning the distribution of the input data in the feature space. On the other hand, it can approximate arbitrary feature distribution by introducing GAN. Concretely, SSAVAE considers the unlabeled data as the problem of the lack of emotional label attributes. The labeled data share the category information with the unlabeled data in the feature space. Several experiments are conducted on the FAU Aibo dataset to evaluate the effectiveness of the algorithm. The results show that the proposed method is superior to other benchmark algorithms and has strong feature learning capabilities.","PeriodicalId":56007,"journal":{"name":"Journal of Cloud Computing-Advances Systems and Applications","volume":"1 1","pages":"275-280"},"PeriodicalIF":4.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88627494","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-01DOI: 10.1109/CSCloud-EdgeCom58631.2023.00049
Yang Yang, Kai Jin, Wei Liang, Yaqin Liu, Yuhui Li, Osama Hosam
Blockchain technology offers unique advantages in terms of decentralization, transparency, and de-anonymization. However, it also poses challenges to user anonymity and data privacy protection. Consequently, researchers have employed advanced cryptographic primitives to enhance the privacy and anonymity of blockchain-based privacy payments, as well as to extend privacy payment methods to more general forms of privacy computing. Nevertheless, relying on high-level cryptographic primitives and emerging technologies, these solutions have proven challenging for academic and industrial personnel to understand and apply. Therefore, we introduce the principle mechanisms of zero-knowledge proofs and homomorphic encryption and their typical algorithms, analyze and summarize recent research in blockchain privacy computing across several dimensions, and briefly present their potential applications.
{"title":"A Review of Blockchain-based Privacy Computing Research","authors":"Yang Yang, Kai Jin, Wei Liang, Yaqin Liu, Yuhui Li, Osama Hosam","doi":"10.1109/CSCloud-EdgeCom58631.2023.00049","DOIUrl":"https://doi.org/10.1109/CSCloud-EdgeCom58631.2023.00049","url":null,"abstract":"Blockchain technology offers unique advantages in terms of decentralization, transparency, and de-anonymization. However, it also poses challenges to user anonymity and data privacy protection. Consequently, researchers have employed advanced cryptographic primitives to enhance the privacy and anonymity of blockchain-based privacy payments, as well as to extend privacy payment methods to more general forms of privacy computing. Nevertheless, relying on high-level cryptographic primitives and emerging technologies, these solutions have proven challenging for academic and industrial personnel to understand and apply. Therefore, we introduce the principle mechanisms of zero-knowledge proofs and homomorphic encryption and their typical algorithms, analyze and summarize recent research in blockchain privacy computing across several dimensions, and briefly present their potential applications.","PeriodicalId":56007,"journal":{"name":"Journal of Cloud Computing-Advances Systems and Applications","volume":"9 1","pages":"241-246"},"PeriodicalIF":4.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84287580","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：