The swarm cluster has two main limitations: 1) the data of database container will lost after the container goes down; 2) the lack of migration ability of database container across the hosts. To tackle these issues, we propose a novel persistence framework in both single database and database cluster. To be specific, we use ceph to provide migrable data volumes, and use two frameworks to migrate container from the perspective of container downtime recovery. By comparing the processing time of downtime database container, the experimental results demonstrate that our proposed method is able to shorten the recovery time of database container and improve the availability of database services.
{"title":"Database Docker persistence Framework based on Swarm and Ceph","authors":"Shaojia Hong, Dong Li, Xiaobing Huang","doi":"10.1145/3341069.3342985","DOIUrl":"https://doi.org/10.1145/3341069.3342985","url":null,"abstract":"The swarm cluster has two main limitations: 1) the data of database container will lost after the container goes down; 2) the lack of migration ability of database container across the hosts. To tackle these issues, we propose a novel persistence framework in both single database and database cluster. To be specific, we use ceph to provide migrable data volumes, and use two frameworks to migrate container from the perspective of container downtime recovery. By comparing the processing time of downtime database container, the experimental results demonstrate that our proposed method is able to shorten the recovery time of database container and improve the availability of database services.","PeriodicalId":411198,"journal":{"name":"Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference","volume":"90 5","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132364773","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Recurrent neural networks (RNN) and convolutional neural networks (CNN) have been extensively used on text classification to capture the local and long-range dependencies. Recent work has demonstrated the superiority of self-attention networks (SAN) owing to their highly parallelizable computation and excellent performance. However, SAN has difficulty capturing meaningful semantic relationships over very long sequences, and the memory requirement grows rapidly in line with the sequence length. To solve these limitations of SAN in processing long document sequence, this paper proposes four novel ideas and build a hierarchical text-label integrated attention network(HLAN). Firstly, a hierarchical architecture is introduced to map the hierarchy of document, which effectively shortens the sequence length of each process. Secondly, the attention weights are calculated in the joint embedding space of text and label. Thirdly, a multi-head soft attention is proposed to compress the sequence encoded by self-attention into a single vector. Finally, a loss term called class loss is given and combined with cross entropy loss. HLAN achieves competitive results over the compared strong baseline methods on 4 out of 5 benchmark datasets, which verifies the effectiveness of HLAN for document classification, in terms of both accuracy and memory requirement.
{"title":"Hierarchical Text-Label Integrated Attention Network for Document Classification","authors":"Changjin Gong, Kaize Shi, Zhendong Niu","doi":"10.1145/3341069.3342987","DOIUrl":"https://doi.org/10.1145/3341069.3342987","url":null,"abstract":"Recurrent neural networks (RNN) and convolutional neural networks (CNN) have been extensively used on text classification to capture the local and long-range dependencies. Recent work has demonstrated the superiority of self-attention networks (SAN) owing to their highly parallelizable computation and excellent performance. However, SAN has difficulty capturing meaningful semantic relationships over very long sequences, and the memory requirement grows rapidly in line with the sequence length. To solve these limitations of SAN in processing long document sequence, this paper proposes four novel ideas and build a hierarchical text-label integrated attention network(HLAN). Firstly, a hierarchical architecture is introduced to map the hierarchy of document, which effectively shortens the sequence length of each process. Secondly, the attention weights are calculated in the joint embedding space of text and label. Thirdly, a multi-head soft attention is proposed to compress the sequence encoded by self-attention into a single vector. Finally, a loss term called class loss is given and combined with cross entropy loss. HLAN achieves competitive results over the compared strong baseline methods on 4 out of 5 benchmark datasets, which verifies the effectiveness of HLAN for document classification, in terms of both accuracy and memory requirement.","PeriodicalId":411198,"journal":{"name":"Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131818048","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
With the rapid popularization of mobile devices, mobile devices such as smart phones have become an indispensable tool in people's daily life. Mobile devices not only bring convenience to human beings, but also bring criminal activities based on mobile devices such as SMS fraud, dissemination of harmful information, virus software and so on. Therefore, digital forensics for mobile devices under IOS operating system is of great significance for combating crime, information security and other issues. This paper first expounds the background and significance of mobile terminal forensics. This paper presents a digital acquisition method based on usbmuxd and iTunes for IOS devices. The method of parsing and storing raw data and restoring file directory are also provided.
{"title":"Digital Forensics Design of IOS Operating System","authors":"Zhendong Liao, Shunxiang Wu, Bin Xi, Fulin Wang, Daodong Ming, Baihua Chen","doi":"10.1145/3341069.3341081","DOIUrl":"https://doi.org/10.1145/3341069.3341081","url":null,"abstract":"With the rapid popularization of mobile devices, mobile devices such as smart phones have become an indispensable tool in people's daily life. Mobile devices not only bring convenience to human beings, but also bring criminal activities based on mobile devices such as SMS fraud, dissemination of harmful information, virus software and so on. Therefore, digital forensics for mobile devices under IOS operating system is of great significance for combating crime, information security and other issues. This paper first expounds the background and significance of mobile terminal forensics. This paper presents a digital acquisition method based on usbmuxd and iTunes for IOS devices. The method of parsing and storing raw data and restoring file directory are also provided.","PeriodicalId":411198,"journal":{"name":"Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference","volume":"130-132 1-3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131660390","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Aiming at the problem of malicious data deletion or tampering in the untrusted server-side storage, an improved PDP (Provable Data Possession) algorithm supporting privacy protection and multiple copies is proposed in the paper. BLS signature and encrypted copies with random mask are adopted in the algorithm. The implementation of the algorithm is described in detail. Experimental results show that the proposed algorithm achieves better performance compared with the MR-PDP algorithm proposed by Cutmola [6]. Finally, the algorithm is applied in a component testing tool to verify the integrity of component source files before component downloading and deployment.
{"title":"An Privacy Preserving and Multi-copy Supporting PDP Algorithm and its Application in Component Testing Tool","authors":"Yanhua Shi, Guozheng Zhang, Shuyu Li","doi":"10.1145/3341069.3342972","DOIUrl":"https://doi.org/10.1145/3341069.3342972","url":null,"abstract":"Aiming at the problem of malicious data deletion or tampering in the untrusted server-side storage, an improved PDP (Provable Data Possession) algorithm supporting privacy protection and multiple copies is proposed in the paper. BLS signature and encrypted copies with random mask are adopted in the algorithm. The implementation of the algorithm is described in detail. Experimental results show that the proposed algorithm achieves better performance compared with the MR-PDP algorithm proposed by Cutmola [6]. Finally, the algorithm is applied in a component testing tool to verify the integrity of component source files before component downloading and deployment.","PeriodicalId":411198,"journal":{"name":"Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126772401","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Lili Liu, E. Tan, Xieping Yin, Yongda Zhen, Z. Cai
Protective coatings are the primary means of protecting marine and offshore structures from corrosion. Coating breakdown and corrosion (CBC) evaluation is the primary method of coating failure management. Evaluation methods can result in unnecessary maintenance costs and a higher risk of failure. To achieve a comprehensive collection of data for CBC assessment, an unmanned arial system (UAS), assisted by the latest technological innovations, will be used to facilitate data collection in inaccessible locations. An image-based CBC assessment system is developed to provide objective assessment of the severity of coating failure. This method is more suitable for inspecting large areas by capturing and analyzing pictures/videos of the target area than the surveyor's existing manual inspection solution. In this paper, deep learning-based object detection in the CBC assessment system has been developed to provide an effective CBC assessment for the marine and offshore industries. This will greatly improve the efficiency and reliability of coating inspection.
{"title":"Deep learning for Coating Condition Assessment with Active perception","authors":"Lili Liu, E. Tan, Xieping Yin, Yongda Zhen, Z. Cai","doi":"10.1145/3341069.3342966","DOIUrl":"https://doi.org/10.1145/3341069.3342966","url":null,"abstract":"Protective coatings are the primary means of protecting marine and offshore structures from corrosion. Coating breakdown and corrosion (CBC) evaluation is the primary method of coating failure management. Evaluation methods can result in unnecessary maintenance costs and a higher risk of failure. To achieve a comprehensive collection of data for CBC assessment, an unmanned arial system (UAS), assisted by the latest technological innovations, will be used to facilitate data collection in inaccessible locations. An image-based CBC assessment system is developed to provide objective assessment of the severity of coating failure. This method is more suitable for inspecting large areas by capturing and analyzing pictures/videos of the target area than the surveyor's existing manual inspection solution. In this paper, deep learning-based object detection in the CBC assessment system has been developed to provide an effective CBC assessment for the marine and offshore industries. This will greatly improve the efficiency and reliability of coating inspection.","PeriodicalId":411198,"journal":{"name":"Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127675131","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Time series processing is a vital issue that is encountered in various fields. However, such data are mostly non-stationary on account of the fact that they are affected by a variety of factors. In this paper, we present a supervised strategy by integrating the iterative filtering (IF) method and convolution neural network (CNN) to automatically extract features of time series, where the IF technique can decompose the raw time series into intrinsic mode functions (IMFs), and then the CNN aims to extract the features from the images constructed by the IMFs under specific task. To illustrate the performance of the proposed strategy, we apply it in one-step and multi-step predictive tasks on the national association of securities dealers automated quotations (NASDAQ) data. Furthermore, we compute the importance of the extracted and raw features by the combined decision trees, such as random forest (RF) and gradient boosted decision trees (GBDT). The results indicate the significant improvement of the proposed strategy.
{"title":"A Strategy Integrating Iterative Filtering and Convolution Neural Network for Time Series Feature Extraction","authors":"Feng Zhou, Liu Jiang","doi":"10.1145/3341069.3342994","DOIUrl":"https://doi.org/10.1145/3341069.3342994","url":null,"abstract":"Time series processing is a vital issue that is encountered in various fields. However, such data are mostly non-stationary on account of the fact that they are affected by a variety of factors. In this paper, we present a supervised strategy by integrating the iterative filtering (IF) method and convolution neural network (CNN) to automatically extract features of time series, where the IF technique can decompose the raw time series into intrinsic mode functions (IMFs), and then the CNN aims to extract the features from the images constructed by the IMFs under specific task. To illustrate the performance of the proposed strategy, we apply it in one-step and multi-step predictive tasks on the national association of securities dealers automated quotations (NASDAQ) data. Furthermore, we compute the importance of the extracted and raw features by the combined decision trees, such as random forest (RF) and gradient boosted decision trees (GBDT). The results indicate the significant improvement of the proposed strategy.","PeriodicalId":411198,"journal":{"name":"Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114681636","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In wireless communication transmission system, it is difficult to obtain the perfect channel state information (CSI) due to the stochastic channel condition and link delay, but most of studies for non-orthogonal multiple access (NOMA) heterogeneous networks always assume that the base station (BS) has perfect CSI. Therefore, in this paper, we study the energy efficient power allocation with a full consideration of users fairness for the NOMA heterogeneous networks based on the imperfect CSI. The probabilistic optimization problem with outage probability is transformed to a non-probabilistic problem via the approximation of inequality. The binary search method is designed to solve the power allocation problem of small cells. The simulation results show that the proposed algorithm with different parameters can greatly improve the energy efficiency of the system.
{"title":"Fair Optimal Power Allocation for Non-orthogonal Multiple Access Heterogeneous Networks","authors":"Xin Song, Li Dong, Lei Qin","doi":"10.1145/3341069.3341088","DOIUrl":"https://doi.org/10.1145/3341069.3341088","url":null,"abstract":"In wireless communication transmission system, it is difficult to obtain the perfect channel state information (CSI) due to the stochastic channel condition and link delay, but most of studies for non-orthogonal multiple access (NOMA) heterogeneous networks always assume that the base station (BS) has perfect CSI. Therefore, in this paper, we study the energy efficient power allocation with a full consideration of users fairness for the NOMA heterogeneous networks based on the imperfect CSI. The probabilistic optimization problem with outage probability is transformed to a non-probabilistic problem via the approximation of inequality. The binary search method is designed to solve the power allocation problem of small cells. The simulation results show that the proposed algorithm with different parameters can greatly improve the energy efficiency of the system.","PeriodicalId":411198,"journal":{"name":"Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128303766","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jianlong Ren, Li Yang, Chun Zuo, Weiyi Kong, Xiaoxiao Ma
Modeling and reasoning about the dialogue history is a main challenge for building a good multi-turn conversational agent. End-to-end memory networks with recurrent or gated architectures have been demonstrated promising for conversation modeling. However, it still suffers from relatively low computational efficiency for its complex architectures and costly strong supervision information or fixed priori knowledge. This paper proposes a multi-head attention based end-to-end approach called multi-attending memory network without additional information or knowledge, which can effectively model and reason about multi-turn history dialogue. Specifically, a parallel multi-head attention mechanism is introduced to model conversational context via attending to different important sections of a full dialog. Thereafter, a stacked architecture with shortcut connections is presented to reason about the memory (the result of context modeling). Experiments on the bAbI-dialog datasets demonstrate the effectiveness of proposed approach.
{"title":"Multi-attending Memory Network for Modeling Multi-turn Dialogue","authors":"Jianlong Ren, Li Yang, Chun Zuo, Weiyi Kong, Xiaoxiao Ma","doi":"10.1145/3341069.3342970","DOIUrl":"https://doi.org/10.1145/3341069.3342970","url":null,"abstract":"Modeling and reasoning about the dialogue history is a main challenge for building a good multi-turn conversational agent. End-to-end memory networks with recurrent or gated architectures have been demonstrated promising for conversation modeling. However, it still suffers from relatively low computational efficiency for its complex architectures and costly strong supervision information or fixed priori knowledge. This paper proposes a multi-head attention based end-to-end approach called multi-attending memory network without additional information or knowledge, which can effectively model and reason about multi-turn history dialogue. Specifically, a parallel multi-head attention mechanism is introduced to model conversational context via attending to different important sections of a full dialog. Thereafter, a stacked architecture with shortcut connections is presented to reason about the memory (the result of context modeling). Experiments on the bAbI-dialog datasets demonstrate the effectiveness of proposed approach.","PeriodicalId":411198,"journal":{"name":"Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115480536","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The technology in the field of multimedia image processing improves every day, but there are still some problems that deserve to be further explored and improved. People like to take and preserve the impressive scenery as an unforgettable memory. However, if the object is moving or the photographer is shaking, the captured image is easily blurred, and this blur is called motion blur. However, deblurring an image without the information of speed and direction of moving objects is still a well-known ill-posed problem. In this paper, we proposed a system to deblur image that can estimate important parameter advance to reduce the complexity of deblurring process. The data of sensor of moving object is collected. The BPN neural network is used to train to classify the speed and direction of the object from the sensor data. After that, we can estimate the speed and direction of objects without other algorithms. With such important parameters, deblurring processing will more efficient.
{"title":"Estimating Parameters for Deblurring in Two-Dimensional Linear Motion","authors":"Chu-Hui Lee, Yong-Jin Zhuo","doi":"10.1145/3341069.3341089","DOIUrl":"https://doi.org/10.1145/3341069.3341089","url":null,"abstract":"The technology in the field of multimedia image processing improves every day, but there are still some problems that deserve to be further explored and improved. People like to take and preserve the impressive scenery as an unforgettable memory. However, if the object is moving or the photographer is shaking, the captured image is easily blurred, and this blur is called motion blur. However, deblurring an image without the information of speed and direction of moving objects is still a well-known ill-posed problem. In this paper, we proposed a system to deblur image that can estimate important parameter advance to reduce the complexity of deblurring process. The data of sensor of moving object is collected. The BPN neural network is used to train to classify the speed and direction of the object from the sensor data. After that, we can estimate the speed and direction of objects without other algorithms. With such important parameters, deblurring processing will more efficient.","PeriodicalId":411198,"journal":{"name":"Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124908188","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The emerging attention based methods are widely used in sentiment classification, achieving the accuracy improvement of sediment classification tasks. However, these methods usually work improperly in the task of film review classification, in which positive and negative comments are often mixed and interpreting the comments from different perspectives may be diametrically opposite sentiments. In this paper, we propose a new attention based neural network architecture based on HAN model where context layer is added. Compared with the HAN, the addition of the context-aspect layer can remove the impact of unimportant sentences and improve the accuracy of sentiment classification. The experiment results on IMDB dataset show that the proposed model outperforms other existing methods, achieving an accuracy improvement of 3.11% as compared to the state-of-the-art method. The experiment results also show that our model has the better accuracy and the lower iteration time, as compared to the baseline model.
{"title":"Text Sentiment Classification Based on Layered Attention Network","authors":"Jinhao Wu, Kai Zheng, Jun Sun","doi":"10.1145/3341069.3342990","DOIUrl":"https://doi.org/10.1145/3341069.3342990","url":null,"abstract":"The emerging attention based methods are widely used in sentiment classification, achieving the accuracy improvement of sediment classification tasks. However, these methods usually work improperly in the task of film review classification, in which positive and negative comments are often mixed and interpreting the comments from different perspectives may be diametrically opposite sentiments. In this paper, we propose a new attention based neural network architecture based on HAN model where context layer is added. Compared with the HAN, the addition of the context-aspect layer can remove the impact of unimportant sentences and improve the accuracy of sentiment classification. The experiment results on IMDB dataset show that the proposed model outperforms other existing methods, achieving an accuracy improvement of 3.11% as compared to the state-of-the-art method. The experiment results also show that our model has the better accuracy and the lower iteration time, as compared to the baseline model.","PeriodicalId":411198,"journal":{"name":"Proceedings of the 2019 3rd High Performance Computing and Cluster Technologies Conference","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125088676","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","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：