Pub Date : 2020-11-01DOI: 10.1109/ICDCS47774.2020.00125
Hao Wu, Wei Liu, Yifan Gong, Jiangming Jin
GPUs have been widely adopted to speedup various throughput-originated applications running on HPC platforms, where typically there are a number of tasks sharing GPUs to maximize GPU utilization. To facilitate GPU sharing, GPU vendors provide tools, allowing multiple processes concurrently to use GPUs. For example, Nvidia provides MPS (Multi-Process Service) managing all GPU processes to achieve high throughput by fully exploiting hardware resources. However, such tool leads to undesired single point of failure for all GPU processes, namely, one process’s exception makes other processes abnormal. In this work, we investigate the seriousness of this GPU process interferences caused by MPS, and propose an approach to address one of these interferences, which takes place during process quitting. By using signal handling and thread synchronization techniques in this approach, GPU processes are able to quit safely without interfering other GPU processes.
{"title":"Safe Process Quitting for GPU Multi-Process Service (MPS)","authors":"Hao Wu, Wei Liu, Yifan Gong, Jiangming Jin","doi":"10.1109/ICDCS47774.2020.00125","DOIUrl":"https://doi.org/10.1109/ICDCS47774.2020.00125","url":null,"abstract":"GPUs have been widely adopted to speedup various throughput-originated applications running on HPC platforms, where typically there are a number of tasks sharing GPUs to maximize GPU utilization. To facilitate GPU sharing, GPU vendors provide tools, allowing multiple processes concurrently to use GPUs. For example, Nvidia provides MPS (Multi-Process Service) managing all GPU processes to achieve high throughput by fully exploiting hardware resources. However, such tool leads to undesired single point of failure for all GPU processes, namely, one process’s exception makes other processes abnormal. In this work, we investigate the seriousness of this GPU process interferences caused by MPS, and propose an approach to address one of these interferences, which takes place during process quitting. By using signal handling and thread synchronization techniques in this approach, GPU processes are able to quit safely without interfering other GPU processes.","PeriodicalId":158630,"journal":{"name":"2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128607914","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}
Pub Date : 2020-11-01DOI: 10.1109/ICDCS47774.2020.00014
K. Ren, Yu Guo, Jiaqi Li, X. Jia, Cong Wang, Yajin Zhou, Sheng Wang, N. Cao, Feifei Li
An encrypted index is a data structure that assisting untrusted servers to provide various query functionalities in the ciphertext domain. Although traditional index designs can prevent servers from directly obtaining plaintexts, the confidentiality of outsourced data could still be compromised by observing the volume of different queries. Recent volume attacks have demonstrated the importance of sealing volume-pattern leakage. To this end, several works are made to design secure indexes with the volume-hiding property. However, prior designs only work for encrypted keyword search. Due to the unpredictable range query results, it is difficult to protect the volume-pattern leakage while achieving efficient range queries.In this paper, for the first time, we define and solve the challenging problem of volume-hiding range queries over encrypted data. Our proposed hybrid index framework, called HybrIDX, allows an untrusted server to efficiently search encrypted data based on order conditions without revealing the exact volume size. It resorts to the trusted hardware techniques to assist range query processing by moving the comparison algorithm to trusted SGX enclaves. To enable volume-hiding data retrieval, we propose to host encrypted results outside the enclave in an encrypted multimaps manner. Apart from this novel hybrid index design, we further customize a bulk refresh mechanism to enable accesspattern obfuscation. We formally analyze the security strengths and complete the prototype implementation. Evaluation results demonstrate the feasibility and practicability of our designs.
{"title":"HybrIDX: New Hybrid Index for Volume-hiding Range Queries in Data Outsourcing Services","authors":"K. Ren, Yu Guo, Jiaqi Li, X. Jia, Cong Wang, Yajin Zhou, Sheng Wang, N. Cao, Feifei Li","doi":"10.1109/ICDCS47774.2020.00014","DOIUrl":"https://doi.org/10.1109/ICDCS47774.2020.00014","url":null,"abstract":"An encrypted index is a data structure that assisting untrusted servers to provide various query functionalities in the ciphertext domain. Although traditional index designs can prevent servers from directly obtaining plaintexts, the confidentiality of outsourced data could still be compromised by observing the volume of different queries. Recent volume attacks have demonstrated the importance of sealing volume-pattern leakage. To this end, several works are made to design secure indexes with the volume-hiding property. However, prior designs only work for encrypted keyword search. Due to the unpredictable range query results, it is difficult to protect the volume-pattern leakage while achieving efficient range queries.In this paper, for the first time, we define and solve the challenging problem of volume-hiding range queries over encrypted data. Our proposed hybrid index framework, called HybrIDX, allows an untrusted server to efficiently search encrypted data based on order conditions without revealing the exact volume size. It resorts to the trusted hardware techniques to assist range query processing by moving the comparison algorithm to trusted SGX enclaves. To enable volume-hiding data retrieval, we propose to host encrypted results outside the enclave in an encrypted multimaps manner. Apart from this novel hybrid index design, we further customize a bulk refresh mechanism to enable accesspattern obfuscation. We formally analyze the security strengths and complete the prototype implementation. Evaluation results demonstrate the feasibility and practicability of our designs.","PeriodicalId":158630,"journal":{"name":"2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129261449","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}
Pub Date : 2020-11-01DOI: 10.1109/ICDCS47774.2020.00111
Yan Li, Bo An, Junming Ma, Donggang Cao, Yasha Wang, Hong Mei
Hyper-parameter tuning (HPT) is crucial for many machine learning (ML) algorithms. But due to the large searching space, HPT is usually time-consuming and resource-intensive. Nowadays, many researchers use public cloud resources to train machine learning models, convenient yet expensive. How to speed up the HPT process while at the same time reduce cost is very important for cloud ML users. In this paper, we propose SpotTune, an approach that exploits transient revocable resources in the public cloud with some tailored strategies to do HPT in a parallel and cost-efficient manner. Orchestrating the HPT process upon transient servers, SpotTune uses two main techniques, fine-grained cost-aware resource provisioning, and ML training trend predicting, to reduce the monetary cost and runtime of HPT processes. Our evaluations show that SpotTune can reduce the cost by up to 90% and achieve a 16.61x performance-cost rate improvement.
{"title":"SpotTune: Leveraging Transient Resources for Cost-efficient Hyper-parameter Tuning in the Public Cloud","authors":"Yan Li, Bo An, Junming Ma, Donggang Cao, Yasha Wang, Hong Mei","doi":"10.1109/ICDCS47774.2020.00111","DOIUrl":"https://doi.org/10.1109/ICDCS47774.2020.00111","url":null,"abstract":"Hyper-parameter tuning (HPT) is crucial for many machine learning (ML) algorithms. But due to the large searching space, HPT is usually time-consuming and resource-intensive. Nowadays, many researchers use public cloud resources to train machine learning models, convenient yet expensive. How to speed up the HPT process while at the same time reduce cost is very important for cloud ML users. In this paper, we propose SpotTune, an approach that exploits transient revocable resources in the public cloud with some tailored strategies to do HPT in a parallel and cost-efficient manner. Orchestrating the HPT process upon transient servers, SpotTune uses two main techniques, fine-grained cost-aware resource provisioning, and ML training trend predicting, to reduce the monetary cost and runtime of HPT processes. Our evaluations show that SpotTune can reduce the cost by up to 90% and achieve a 16.61x performance-cost rate improvement.","PeriodicalId":158630,"journal":{"name":"2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128252721","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}
Pub Date : 2020-11-01DOI: 10.1109/ICDCS47774.2020.00178
Rui Liang, Shengyong Xu
In recent years, 3D reconstruction technology has developed rapidly. It is a promising field to apply 3D reconstruction technology to the measurement of plant configuration parameters. The main content of this paper is the 3D reconstruction technology for rape roots and the measurement methods for their key traits. Firstly, we set up a set of low-cost image sequence acquisition device of rape roots. We collected image data with common consumption level camera and used the method of SfM to carry out 3D reconstruction of rape roots. Then we proposed a series of algorithms to measure the surface area, volume, number of primary lateral roots and length of taproot based on the huge point cloud data obtained from 3D reconstruction. Finally, we designed a set of nondestructive measurement system for key traits of rape roots. The total volume of root, the number of primary lateral roots and the length of taproot were measured manually. Compared with the results of manual measurement, the accuracy of the main algorithm proposed in this paper is not less than 95%. Our contribution is to provide a 3D reconstruction method that is easy to operate, and to provide a high-precision measurement method for the key traits of rape roots, which has an important value for quantitative analysis of rape roots phenotype.
{"title":"Three-Dimensional Reconstruction and Phenotype Nondestructive Measurement Technology for Rape Roots","authors":"Rui Liang, Shengyong Xu","doi":"10.1109/ICDCS47774.2020.00178","DOIUrl":"https://doi.org/10.1109/ICDCS47774.2020.00178","url":null,"abstract":"In recent years, 3D reconstruction technology has developed rapidly. It is a promising field to apply 3D reconstruction technology to the measurement of plant configuration parameters. The main content of this paper is the 3D reconstruction technology for rape roots and the measurement methods for their key traits. Firstly, we set up a set of low-cost image sequence acquisition device of rape roots. We collected image data with common consumption level camera and used the method of SfM to carry out 3D reconstruction of rape roots. Then we proposed a series of algorithms to measure the surface area, volume, number of primary lateral roots and length of taproot based on the huge point cloud data obtained from 3D reconstruction. Finally, we designed a set of nondestructive measurement system for key traits of rape roots. The total volume of root, the number of primary lateral roots and the length of taproot were measured manually. Compared with the results of manual measurement, the accuracy of the main algorithm proposed in this paper is not less than 95%. Our contribution is to provide a 3D reconstruction method that is easy to operate, and to provide a high-precision measurement method for the key traits of rape roots, which has an important value for quantitative analysis of rape roots phenotype.","PeriodicalId":158630,"journal":{"name":"2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124363059","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}
Pub Date : 2020-11-01DOI: 10.1109/ICDCS47774.2020.00035
Anirudh Ganji, Anandeshwar Singh, Muhammad Shahzad
The switch fabrics of today’s data centers carry traffic controlled by a variety of TCP congestion control algorithms. This leads us to ask: how does the coexistence of multiple variants of TCP on shared switch fabric impacts the performance achieved by different applications in data centers? To answer this question, we conducted an extensive set of experiments with coexisting TCP variants on Leaf-Spine and Fat-Tree switch fabrics. We executed common data center workloads, which include streaming, MapReduce, and storage workloads, using four commonly used TCP variants, namely BBR, DCTCP, CUBIC, and New Reno. We also extensively executed iPerf workloads using these 4 TCP variants to purely study the impact of the coexistence of TCP variants on each other’s performance without incorporating the network behavior of the application layer. Our experiments resulted in a large set of network traces comprised of 160 billion packets (we will release these traces after publication of this work). We present comprehensive observations from these traces that have important implications in ensuring optimal utilization of data center switch fabric and in meeting the network performance needs of application layer workloads.
{"title":"Characterizing the Impact of TCP Coexistence in Data Center Networks","authors":"Anirudh Ganji, Anandeshwar Singh, Muhammad Shahzad","doi":"10.1109/ICDCS47774.2020.00035","DOIUrl":"https://doi.org/10.1109/ICDCS47774.2020.00035","url":null,"abstract":"The switch fabrics of today’s data centers carry traffic controlled by a variety of TCP congestion control algorithms. This leads us to ask: how does the coexistence of multiple variants of TCP on shared switch fabric impacts the performance achieved by different applications in data centers? To answer this question, we conducted an extensive set of experiments with coexisting TCP variants on Leaf-Spine and Fat-Tree switch fabrics. We executed common data center workloads, which include streaming, MapReduce, and storage workloads, using four commonly used TCP variants, namely BBR, DCTCP, CUBIC, and New Reno. We also extensively executed iPerf workloads using these 4 TCP variants to purely study the impact of the coexistence of TCP variants on each other’s performance without incorporating the network behavior of the application layer. Our experiments resulted in a large set of network traces comprised of 160 billion packets (we will release these traces after publication of this work). We present comprehensive observations from these traces that have important implications in ensuring optimal utilization of data center switch fabric and in meeting the network performance needs of application layer workloads.","PeriodicalId":158630,"journal":{"name":"2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127823900","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}
Pub Date : 2020-11-01DOI: 10.1109/ICDCS47774.2020.00183
Kuldeep Sharma, N. Ramakrishnan, Alok Prakash, S. Lam, T. Srikanthan
Pruning of channels in trained deep neural networks has been widely used to implement efficient DNNs that can be deployed on embedded/mobile devices. Majority of existing techniques employ criteria-based sorting of the channels to preserve salient channels during pruning as well as to automatically determine the pruned network architecture. However, recent studies on widely used DNNs, such as VGG-16, have shown that selecting and preserving salient channels using pruning criteria is not necessary since the plasticity of the network allows the accuracy to be recovered through fine-tuning. In this work, we further explore the value of the ranking criteria in pruning to show that if channels are removed gradually and iteratively, alternating with fine-tuning on the target dataset, ranking criteria are indeed not necessary to select redundant channels. Experimental results confirm that even a random selection of channels for pruning leads to similar performance (accuracy). In addition, we demonstrate that even a simple pruning technique that uniformly removes channels from all layers in the network, performs similar to existing ranking criteria-based approaches, while leading to lower inference time (GFLOPs). Our extensive evaluations include the context of embedded implementations of DNNs - specifically, on small networks such as SqueezeNet and at aggressive pruning percentages. We leverage these insights, to propose a GFLOPs-aware iterative pruning strategy that does not rely on any ranking criteria and yet can further lead to lower inference time by 15% without sacrificing accuracy.
{"title":"Evaluating the Merits of Ranking in Structured Network Pruning","authors":"Kuldeep Sharma, N. Ramakrishnan, Alok Prakash, S. Lam, T. Srikanthan","doi":"10.1109/ICDCS47774.2020.00183","DOIUrl":"https://doi.org/10.1109/ICDCS47774.2020.00183","url":null,"abstract":"Pruning of channels in trained deep neural networks has been widely used to implement efficient DNNs that can be deployed on embedded/mobile devices. Majority of existing techniques employ criteria-based sorting of the channels to preserve salient channels during pruning as well as to automatically determine the pruned network architecture. However, recent studies on widely used DNNs, such as VGG-16, have shown that selecting and preserving salient channels using pruning criteria is not necessary since the plasticity of the network allows the accuracy to be recovered through fine-tuning. In this work, we further explore the value of the ranking criteria in pruning to show that if channels are removed gradually and iteratively, alternating with fine-tuning on the target dataset, ranking criteria are indeed not necessary to select redundant channels. Experimental results confirm that even a random selection of channels for pruning leads to similar performance (accuracy). In addition, we demonstrate that even a simple pruning technique that uniformly removes channels from all layers in the network, performs similar to existing ranking criteria-based approaches, while leading to lower inference time (GFLOPs). Our extensive evaluations include the context of embedded implementations of DNNs - specifically, on small networks such as SqueezeNet and at aggressive pruning percentages. We leverage these insights, to propose a GFLOPs-aware iterative pruning strategy that does not rely on any ranking criteria and yet can further lead to lower inference time by 15% without sacrificing accuracy.","PeriodicalId":158630,"journal":{"name":"2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127179275","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}
Pub Date : 2020-11-01DOI: 10.1109/ICDCS47774.2020.00077
Qing Li, Nanyang Huang, Yong Jiang, R. Sinnott, Mingwei Xu
Data plane programming languages enable administrators of Software-Defined Networks (SDNs) to perform fine-grained flow control by compiling high-level policies into low-level rules and deploying rules in the data plane. However, it is difficult to scale the data plane with the dynamics of network traffic and the limited storage space of switches. In this paper, we propose a lazy OpenFlow Rule Placement (ORP) framework to enforce control polices and scale the SDN data plane by placing and reusing wildcard rules. We provide an offline rule placement scheme to meet performance objectives under real-world constraints. To handle dynamic traffic and perform incremental rule updates, we design an online matching rule deployment algorithm to place rules in polynomial time and prove it to be conditionally-optimal. Furthermore, to address the rule dependency problem during online rule placement, we extend the algorithm to deploy dependent rules and present lightweight heuristics to guarantee the fast reaction to the new flows. Extensive experiments are conducted on diverse network topologies and datasets to show that the lazy ORP framework significantly reduces the storage cost, improves data plane scalability and is flexible enough to accomplish different optimization goals.
{"title":"Scale the Data Plane of Software-Defined Networks: a Lazy Rule Placement Approach","authors":"Qing Li, Nanyang Huang, Yong Jiang, R. Sinnott, Mingwei Xu","doi":"10.1109/ICDCS47774.2020.00077","DOIUrl":"https://doi.org/10.1109/ICDCS47774.2020.00077","url":null,"abstract":"Data plane programming languages enable administrators of Software-Defined Networks (SDNs) to perform fine-grained flow control by compiling high-level policies into low-level rules and deploying rules in the data plane. However, it is difficult to scale the data plane with the dynamics of network traffic and the limited storage space of switches. In this paper, we propose a lazy OpenFlow Rule Placement (ORP) framework to enforce control polices and scale the SDN data plane by placing and reusing wildcard rules. We provide an offline rule placement scheme to meet performance objectives under real-world constraints. To handle dynamic traffic and perform incremental rule updates, we design an online matching rule deployment algorithm to place rules in polynomial time and prove it to be conditionally-optimal. Furthermore, to address the rule dependency problem during online rule placement, we extend the algorithm to deploy dependent rules and present lightweight heuristics to guarantee the fast reaction to the new flows. Extensive experiments are conducted on diverse network topologies and datasets to show that the lazy ORP framework significantly reduces the storage cost, improves data plane scalability and is flexible enough to accomplish different optimization goals.","PeriodicalId":158630,"journal":{"name":"2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126561714","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}
Pub Date : 2020-11-01DOI: 10.1109/ICDCS47774.2020.00079
Chi Lin, Tingting Xu, Jie Xiong, Fenglong Ma, Lei Wang, Guowei Wu
Handwriting recognition system provides people a convenient and alternative way for writing in the air with fingers rather than typing keyboards. For people with blurred vision and patients with generalized hand neurological disease, writing in the air is particularly attracting due to the small input screen of smartphones and smartwatches. Existing recognition systems still face drawbacks such as requiring to wear dedicated devices, relatively low accuracy and infeasible for cross domain identification, which greatly limit the usability of these systems. To address these issues, we propose WiWrite, an accurate device-free handwriting recognition system which allows writing in the air without a need of attaching any device to the user. Specifically, we use Commercial Off-The-Shelf (COTS) WiFi hardware to achieve fine-grained finger tracking. We develop a CSI division scheme to process the noisy raw WiFi channel state information (CSI), which stabilizes the CSI phase and reduces the noise of the CSI amplitude. To automatically retain low noise data for identification, we propose a self-paced dense convolutional network (SPDCN), which consists of the self-paced loss function based on a modified convolutional neural network, together with a dense convolutional network. Comprehensive experiments are conducted to show the merits of WiWrite, revealing that, the recognition accuracies for the same-size input and different-size input are 93.6% and 89.0%, respectively. Moreover, WiWrite can achieve a one-fit-for-all recognition regardless of environment diversities.
{"title":"WiWrite: An Accurate Device-Free Handwriting Recognition System with COTS WiFi","authors":"Chi Lin, Tingting Xu, Jie Xiong, Fenglong Ma, Lei Wang, Guowei Wu","doi":"10.1109/ICDCS47774.2020.00079","DOIUrl":"https://doi.org/10.1109/ICDCS47774.2020.00079","url":null,"abstract":"Handwriting recognition system provides people a convenient and alternative way for writing in the air with fingers rather than typing keyboards. For people with blurred vision and patients with generalized hand neurological disease, writing in the air is particularly attracting due to the small input screen of smartphones and smartwatches. Existing recognition systems still face drawbacks such as requiring to wear dedicated devices, relatively low accuracy and infeasible for cross domain identification, which greatly limit the usability of these systems. To address these issues, we propose WiWrite, an accurate device-free handwriting recognition system which allows writing in the air without a need of attaching any device to the user. Specifically, we use Commercial Off-The-Shelf (COTS) WiFi hardware to achieve fine-grained finger tracking. We develop a CSI division scheme to process the noisy raw WiFi channel state information (CSI), which stabilizes the CSI phase and reduces the noise of the CSI amplitude. To automatically retain low noise data for identification, we propose a self-paced dense convolutional network (SPDCN), which consists of the self-paced loss function based on a modified convolutional neural network, together with a dense convolutional network. Comprehensive experiments are conducted to show the merits of WiWrite, revealing that, the recognition accuracies for the same-size input and different-size input are 93.6% and 89.0%, respectively. Moreover, WiWrite can achieve a one-fit-for-all recognition regardless of environment diversities.","PeriodicalId":158630,"journal":{"name":"2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129767944","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}
Pub Date : 2020-11-01DOI: 10.1109/ICDCS47774.2020.00086
Liyao Xiang, Lingdong Wang, Shufan Wang, Baochun Li
Machine learning algorithms have been widely deployed on decentralized systems so that users with private, local data can jointly contribute to a better generalized model. One promising approach is Aggregation of Teacher Ensembles, which transfers knowledge of locally trained models to a global one without releasing any private data. However, previous methods largely focus on privately aggregating the local results without concerning their validity, which easily leads to erroneous aggregation results especially when data is unbalanced across different users. Hence, we propose a private consensus protocol — which reveals nothing else but the label with the highest votes, in the condition that the number of votes exceeds a given threshold. The purpose is to filter out undesired aggregation results that could hurt the aggregator model performance. Our protocol also guarantees differential privacy such that any adversary with auxiliary information cannot gain any additional knowledge from the results. We show that with our protocol, we achieve the same privacy level with an improved accuracy compared to previous works.
{"title":"Achieving Consensus in Privacy-Preserving Decentralized Learning","authors":"Liyao Xiang, Lingdong Wang, Shufan Wang, Baochun Li","doi":"10.1109/ICDCS47774.2020.00086","DOIUrl":"https://doi.org/10.1109/ICDCS47774.2020.00086","url":null,"abstract":"Machine learning algorithms have been widely deployed on decentralized systems so that users with private, local data can jointly contribute to a better generalized model. One promising approach is Aggregation of Teacher Ensembles, which transfers knowledge of locally trained models to a global one without releasing any private data. However, previous methods largely focus on privately aggregating the local results without concerning their validity, which easily leads to erroneous aggregation results especially when data is unbalanced across different users. Hence, we propose a private consensus protocol — which reveals nothing else but the label with the highest votes, in the condition that the number of votes exceeds a given threshold. The purpose is to filter out undesired aggregation results that could hurt the aggregator model performance. Our protocol also guarantees differential privacy such that any adversary with auxiliary information cannot gain any additional knowledge from the results. We show that with our protocol, we achieve the same privacy level with an improved accuracy compared to previous works.","PeriodicalId":158630,"journal":{"name":"2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131207105","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}
Pub Date : 2020-11-01DOI: 10.1109/ICDCS47774.2020.00179
Deji Zhao, Bo Ning, Chao Yang
In order to solve the challenges brought by the operation and maintenance of power system in the era of big data, APM (Application Performance Management) system is introduced, which can monitor the operation of software and hardware system, show the health of system operation, and find the performance bottleneck. On the Hadoop platform, a big data deep mining and analysis platform based on map / reduce mode is built, integrating regression analysis, association analysis, data classification, data clustering, text mining, web mining and other data mining algorithms. It can complete 100TB level data retrieval in 30s, and then analyze; the system monitoring server can run stably in a cluster of 256 nodes. The use of APM system can prevent performance bottlenecks, greatly reduce the response time of performance problem processing, and quickly locate the location of performance problems, so as to ensure higher availability and stability of information system.
{"title":"Application research on application performance management system in big data of power grid","authors":"Deji Zhao, Bo Ning, Chao Yang","doi":"10.1109/ICDCS47774.2020.00179","DOIUrl":"https://doi.org/10.1109/ICDCS47774.2020.00179","url":null,"abstract":"In order to solve the challenges brought by the operation and maintenance of power system in the era of big data, APM (Application Performance Management) system is introduced, which can monitor the operation of software and hardware system, show the health of system operation, and find the performance bottleneck. On the Hadoop platform, a big data deep mining and analysis platform based on map / reduce mode is built, integrating regression analysis, association analysis, data classification, data clustering, text mining, web mining and other data mining algorithms. It can complete 100TB level data retrieval in 30s, and then analyze; the system monitoring server can run stably in a cluster of 256 nodes. The use of APM system can prevent performance bottlenecks, greatly reduce the response time of performance problem processing, and quickly locate the location of performance problems, so as to ensure higher availability and stability of information system.","PeriodicalId":158630,"journal":{"name":"2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133410825","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：