Pub Date : 2018-11-01DOI: 10.1109/ICSAI.2018.8599432
Ruifang Li, L. Jin
In traditional cognitive wireless network, most studies on spectrum allocation are on the basis of static network topology. However, the vehicles in the cognitive vehicular network have high-speed mobility and the network topology changes frequently, which makes spectrum allocation more challenging. In this paper, the above factors are considered and a connection between the remaining available time of the primary user and the time required by the cognitive vehicle is established in our spectrum allocation model. To maximize network throughput under the heterogeneous spectrum environment, a rapid convergence algorithm that adapts to a dynamic cognitive vehicular network environment for solving this problem is necessary. Therefore, the improved adaptive binary cuckoo search (IABCS) algorithm that incorporates the simplex method into the adaptive binary cuckoo algorithm is proposed. The experimental results indicate that comparing with the original standard cuckoo search $(CS)$ algorithm and the improved particle swarm optimization (PSO) algorithm, the spectrum allocation method based on the improved adaptive cuckoo algorithm converges faster and achieves higher throughput.
{"title":"Improved Cuckoo Algorithm for Spectrum Allocation in Cognitive Vehicular Network","authors":"Ruifang Li, L. Jin","doi":"10.1109/ICSAI.2018.8599432","DOIUrl":"https://doi.org/10.1109/ICSAI.2018.8599432","url":null,"abstract":"In traditional cognitive wireless network, most studies on spectrum allocation are on the basis of static network topology. However, the vehicles in the cognitive vehicular network have high-speed mobility and the network topology changes frequently, which makes spectrum allocation more challenging. In this paper, the above factors are considered and a connection between the remaining available time of the primary user and the time required by the cognitive vehicle is established in our spectrum allocation model. To maximize network throughput under the heterogeneous spectrum environment, a rapid convergence algorithm that adapts to a dynamic cognitive vehicular network environment for solving this problem is necessary. Therefore, the improved adaptive binary cuckoo search (IABCS) algorithm that incorporates the simplex method into the adaptive binary cuckoo algorithm is proposed. The experimental results indicate that comparing with the original standard cuckoo search $(CS)$ algorithm and the improved particle swarm optimization (PSO) algorithm, the spectrum allocation method based on the improved adaptive cuckoo algorithm converges faster and achieves higher throughput.","PeriodicalId":375852,"journal":{"name":"2018 5th International Conference on Systems and Informatics (ICSAI)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115471700","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 : 2018-11-01DOI: 10.1109/ICSAI.2018.8599288
Youmin Zhang, Li Liu, Shun Fu, Fujin Zhong
Entity alignment across knowledge graphs is an important task in web mining. The aligned entities can be used for transferring knowledge across knowledge graphs and benefit several tasks such as cross-lingual knowledge graph construction and knowledge reasoning. This paper propose a representation learning based algorithm for embedding knowledge graph and aligning entities. In particular, considering the multi-type relations in knowledge graph, we select the alignment-task driven representative relations based on the pre-aligned entity pairs. With the help of selected relations, we embed the entities across networks into a common space by modeling entities’ head/tail are with corresponding context vectors. For entity alignment task, pre-aligned entities are adopted to facilitate the transfer of context information across the knowledges graphs. Through this way, the problem of entity embedding and alignment can be solved simultaneously under a unified framework.. Extensive experiments on two multi-lingual knowledge graphs demonstrate the effectiveness of the proposed model comparing with several state-of-the-art models.
{"title":"Entity Alignment Across Knowledge Graphs Based on Representative Relations Selection","authors":"Youmin Zhang, Li Liu, Shun Fu, Fujin Zhong","doi":"10.1109/ICSAI.2018.8599288","DOIUrl":"https://doi.org/10.1109/ICSAI.2018.8599288","url":null,"abstract":"Entity alignment across knowledge graphs is an important task in web mining. The aligned entities can be used for transferring knowledge across knowledge graphs and benefit several tasks such as cross-lingual knowledge graph construction and knowledge reasoning. This paper propose a representation learning based algorithm for embedding knowledge graph and aligning entities. In particular, considering the multi-type relations in knowledge graph, we select the alignment-task driven representative relations based on the pre-aligned entity pairs. With the help of selected relations, we embed the entities across networks into a common space by modeling entities’ head/tail are with corresponding context vectors. For entity alignment task, pre-aligned entities are adopted to facilitate the transfer of context information across the knowledges graphs. Through this way, the problem of entity embedding and alignment can be solved simultaneously under a unified framework.. Extensive experiments on two multi-lingual knowledge graphs demonstrate the effectiveness of the proposed model comparing with several state-of-the-art models.","PeriodicalId":375852,"journal":{"name":"2018 5th International Conference on Systems and Informatics (ICSAI)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123640822","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 : 2018-11-01DOI: 10.1109/ICSAI.2018.8599296
Tao Gao, Xiu-rong Ma, Ming-xin Liu
In this paper, we propose a weighted hard-reliability based one step majority-logic decoding algorithm for NON-Binary Low-Density Parity-Check (NB-LDPC) codes. To improve the information reliable of check nodes and the use efficiency of receive message, a weight reliability message method is proposed where only the weight values generated in the decoding initialization are reserved for the iterate decoding process. We also propose a new message reliability updating rule for each iterate decoding, in which only the unreliable variable nodes are updated. Simulation results show that our proposed weighted iterative hard-reliability (WIHRB) algorithm significantly improves the error-floor performance compared to the conventional truncate iterative hard-reliability (TIHRB) algorithms.
{"title":"Weighted Hard-Reliability Decoding Method for Non-binary LDPC Codes","authors":"Tao Gao, Xiu-rong Ma, Ming-xin Liu","doi":"10.1109/ICSAI.2018.8599296","DOIUrl":"https://doi.org/10.1109/ICSAI.2018.8599296","url":null,"abstract":"In this paper, we propose a weighted hard-reliability based one step majority-logic decoding algorithm for NON-Binary Low-Density Parity-Check (NB-LDPC) codes. To improve the information reliable of check nodes and the use efficiency of receive message, a weight reliability message method is proposed where only the weight values generated in the decoding initialization are reserved for the iterate decoding process. We also propose a new message reliability updating rule for each iterate decoding, in which only the unreliable variable nodes are updated. Simulation results show that our proposed weighted iterative hard-reliability (WIHRB) algorithm significantly improves the error-floor performance compared to the conventional truncate iterative hard-reliability (TIHRB) algorithms.","PeriodicalId":375852,"journal":{"name":"2018 5th International Conference on Systems and Informatics (ICSAI)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126032438","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 : 2018-11-01DOI: 10.1109/ICSAI.2018.8599506
Yuanyuan Sun, Cong Jin, Wei Zhao, Nansu Wang
In order to solve the problems of real-time beat tracking, such as the uncertainty of real beat value, the difficulty of getting close to people’s perception of music and the position of beat according to people’s feelings, the fact that most data sets are private and the amount of data is small, which affects the accuracy of experimental results, a real-time beat tracking method based on lstm neural network is proposed, which abandons the traditional idea of beat tracking to determine the position of beat, divides the beat into five levels according to the degree of strength, and then trains the beat information by using lstm network. Experiments show that the system functions well and the accuracy of the training results is guaranteed to reach 0.946.
{"title":"Design of real-time rhythm tracking system based on neural network","authors":"Yuanyuan Sun, Cong Jin, Wei Zhao, Nansu Wang","doi":"10.1109/ICSAI.2018.8599506","DOIUrl":"https://doi.org/10.1109/ICSAI.2018.8599506","url":null,"abstract":"In order to solve the problems of real-time beat tracking, such as the uncertainty of real beat value, the difficulty of getting close to people’s perception of music and the position of beat according to people’s feelings, the fact that most data sets are private and the amount of data is small, which affects the accuracy of experimental results, a real-time beat tracking method based on lstm neural network is proposed, which abandons the traditional idea of beat tracking to determine the position of beat, divides the beat into five levels according to the degree of strength, and then trains the beat information by using lstm network. Experiments show that the system functions well and the accuracy of the training results is guaranteed to reach 0.946.","PeriodicalId":375852,"journal":{"name":"2018 5th International Conference on Systems and Informatics (ICSAI)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121606889","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 : 2018-11-01DOI: 10.1109/ICSAI.2018.8599297
Yun Liu, Buyang Cao, Yahui Liu
The prediction of the faulty rate of a cellphone is essential for the supply chain management system of spare parts. However, the fault rate of the mobile is affected by many factors that makes it difficult to predict. In this work, some new concepts for prediction of faulty rate based on grey model theory such as grey fault rate and grey model fault count are proposed. It is found that the grey fault rate is consistent with the bathtub curve that widely applied in the reliability engineering. The grey model theory is utilized to solve the problem of random individual fault affecting the prediction negatively. The characteristic value of the grey failure rate is defined to describe the fault rate for certain phones’ models. We develop the method to predict the fault of a new phone model based on the data of certain old phone models and their grey failure rate. The proposed method is applied to fault rate prediction of two cellphone models that results with the prediction deviation about 2% over 3 years.
{"title":"The Prediction of Cellphones’ Fault Rates with Grey Models","authors":"Yun Liu, Buyang Cao, Yahui Liu","doi":"10.1109/ICSAI.2018.8599297","DOIUrl":"https://doi.org/10.1109/ICSAI.2018.8599297","url":null,"abstract":"The prediction of the faulty rate of a cellphone is essential for the supply chain management system of spare parts. However, the fault rate of the mobile is affected by many factors that makes it difficult to predict. In this work, some new concepts for prediction of faulty rate based on grey model theory such as grey fault rate and grey model fault count are proposed. It is found that the grey fault rate is consistent with the bathtub curve that widely applied in the reliability engineering. The grey model theory is utilized to solve the problem of random individual fault affecting the prediction negatively. The characteristic value of the grey failure rate is defined to describe the fault rate for certain phones’ models. We develop the method to predict the fault of a new phone model based on the data of certain old phone models and their grey failure rate. The proposed method is applied to fault rate prediction of two cellphone models that results with the prediction deviation about 2% over 3 years.","PeriodicalId":375852,"journal":{"name":"2018 5th International Conference on Systems and Informatics (ICSAI)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116337063","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 : 2018-11-01DOI: 10.1109/ICSAI.2018.8599504
Jia Cui, Lei Shi, Juan Li, Zhaohui Liu
With rapid development of network technologies, the data accessing paradigm has been transferred from disk-oriented to “on-the-fly” data stream. The string similarity query on data stream has a broad prospect of application, especially in information security area and network monitoring. Due to the characteristics of stream and limitations of computing resources, the current methods based on static dataset cannot support stream efficiently. To solve these challenges, a framework named F2SCQ (framework of string similarity continuous query) based on filtering and verifying approach is pro-posed. It adopts basic window mechanism to update the sliding window, and the improved asymmetric signature (IAS) scheme to extract signature is proposed. Moreover two new filtering algorithms: Pre-Prune Filtering (PPF) and Count Filtering on Stream (CFS) are proposed. The experiments show that F2SCQ achieves high performance over high rates data stream. Compared to q-gram and asymmetric signature scheme, IAS achieves 50% and 20% faster extraction speed and 45% and 9% less storage overhead. The proposed filtering algorithm also achieves faster filtering speed and generates fewer candidates. F2SCQ minimizes the time and space complexity.
{"title":"An Efficient Framework for String Similarity Continuous Query on Data Stream","authors":"Jia Cui, Lei Shi, Juan Li, Zhaohui Liu","doi":"10.1109/ICSAI.2018.8599504","DOIUrl":"https://doi.org/10.1109/ICSAI.2018.8599504","url":null,"abstract":"With rapid development of network technologies, the data accessing paradigm has been transferred from disk-oriented to “on-the-fly” data stream. The string similarity query on data stream has a broad prospect of application, especially in information security area and network monitoring. Due to the characteristics of stream and limitations of computing resources, the current methods based on static dataset cannot support stream efficiently. To solve these challenges, a framework named F2SCQ (framework of string similarity continuous query) based on filtering and verifying approach is pro-posed. It adopts basic window mechanism to update the sliding window, and the improved asymmetric signature (IAS) scheme to extract signature is proposed. Moreover two new filtering algorithms: Pre-Prune Filtering (PPF) and Count Filtering on Stream (CFS) are proposed. The experiments show that F2SCQ achieves high performance over high rates data stream. Compared to q-gram and asymmetric signature scheme, IAS achieves 50% and 20% faster extraction speed and 45% and 9% less storage overhead. The proposed filtering algorithm also achieves faster filtering speed and generates fewer candidates. F2SCQ minimizes the time and space complexity.","PeriodicalId":375852,"journal":{"name":"2018 5th International Conference on Systems and Informatics (ICSAI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126086449","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 : 2018-11-01DOI: 10.1109/ICSAI.2018.8599447
Min Xiao, Mingzi Xiao, Jing Liang, Yan Shi
this paper introduced a temperature controlling system designed for agricultural greenhouses by making use of the ZigBee technology. DS18B20, the in-line temperature sensor was used for temperature collection, which can be transformed into a digit directly. In wireless communication, ZigBee, the protocol stack was used for relevant modification on the application layer so as to implement wireless communication over the ZigBee protocol. The whole temperature collection process and wireless communication were all completed by the CC2530 functional node module by IAR Embedded Workbench, which was an integrated development environment. One functional node of CC2530 was designed to collect temperature and send it as a terminal node. The other functional node of CC2530 was responsible for receiving temperature as a coordinator and transmit the temperature to the STM32F103RBT6 development board by the serial port communication technology. The collected temperature can be shown on the 3.2-inch TFT LCD screen and the display drive can be completed on the STM32F103RBT6 board, on which the rotation of the stepping motor can be driven. The whole development process was implemented by the MDK software developed by Keil. The data can be transmitted between the CC2530 functional node module and the STM32F103RBT6 development board by the serial communication technology. Ultimately, the system test indicated that data can be transmitted correctly and the transmission is stable so that the greenhouse management requirement can be met.
{"title":"ZigBee-based Temperature Controlling System for Agricultural Greenhouses","authors":"Min Xiao, Mingzi Xiao, Jing Liang, Yan Shi","doi":"10.1109/ICSAI.2018.8599447","DOIUrl":"https://doi.org/10.1109/ICSAI.2018.8599447","url":null,"abstract":"this paper introduced a temperature controlling system designed for agricultural greenhouses by making use of the ZigBee technology. DS18B20, the in-line temperature sensor was used for temperature collection, which can be transformed into a digit directly. In wireless communication, ZigBee, the protocol stack was used for relevant modification on the application layer so as to implement wireless communication over the ZigBee protocol. The whole temperature collection process and wireless communication were all completed by the CC2530 functional node module by IAR Embedded Workbench, which was an integrated development environment. One functional node of CC2530 was designed to collect temperature and send it as a terminal node. The other functional node of CC2530 was responsible for receiving temperature as a coordinator and transmit the temperature to the STM32F103RBT6 development board by the serial port communication technology. The collected temperature can be shown on the 3.2-inch TFT LCD screen and the display drive can be completed on the STM32F103RBT6 board, on which the rotation of the stepping motor can be driven. The whole development process was implemented by the MDK software developed by Keil. The data can be transmitted between the CC2530 functional node module and the STM32F103RBT6 development board by the serial communication technology. Ultimately, the system test indicated that data can be transmitted correctly and the transmission is stable so that the greenhouse management requirement can be met.","PeriodicalId":375852,"journal":{"name":"2018 5th International Conference on Systems and Informatics (ICSAI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125442875","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 : 2018-11-01DOI: 10.1109/ICSAI.2018.8599287
Weizhi Ying, Bin Sun, Aoyun Shen, Haifeng Xu, Liangyu Zhong
This paper aims to establish a nonlinear speeding optimizing model for minimizing the fuel consumption. Considering the fuel consumption of the vessel both in sailing and mooring, the traditional speeding optimizing model with fuel consumption only in sailing consideration is improved. Not only the relationship between fuel consumption and speed in sailing is fitted by a power function, but also a linear function was used to fit the relationship between fuel consumption and time in mooring. Based on the two functions above, a new speeding calculating formula which is more practical is proposed. The simulation experiments prove the speeding optimizing model and formula proposed can reduce the fuel consumption and emission more effectively.
{"title":"Speeding optimization considering the fuel consumption in the mooring period","authors":"Weizhi Ying, Bin Sun, Aoyun Shen, Haifeng Xu, Liangyu Zhong","doi":"10.1109/ICSAI.2018.8599287","DOIUrl":"https://doi.org/10.1109/ICSAI.2018.8599287","url":null,"abstract":"This paper aims to establish a nonlinear speeding optimizing model for minimizing the fuel consumption. Considering the fuel consumption of the vessel both in sailing and mooring, the traditional speeding optimizing model with fuel consumption only in sailing consideration is improved. Not only the relationship between fuel consumption and speed in sailing is fitted by a power function, but also a linear function was used to fit the relationship between fuel consumption and time in mooring. Based on the two functions above, a new speeding calculating formula which is more practical is proposed. The simulation experiments prove the speeding optimizing model and formula proposed can reduce the fuel consumption and emission more effectively.","PeriodicalId":375852,"journal":{"name":"2018 5th International Conference on Systems and Informatics (ICSAI)","volume":"239 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126814110","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 : 2018-11-01DOI: 10.1109/ICSAI.2018.8599476
Haizhou Du, Keke Zhang, Zhenchen Yang
In recent years, big data analytics frameworks spring up rapidly. Meanwhile, it has become routine for large volumes of data to be generated, stored, and processed across geographically distributed datac enters. Network congestion generated by data transfers between networks becomes a major bottleneck to the overall performance of the system in a geo-distributed environment. Many existing methods usually process network congestion after they occurs, which does not solve the problem fundamentally. In this paper, we focus on the problem of predicting and avoiding network congestion in advance in a geo-distributed environment on Apache Spark, in terms of their job completion times. We formulate this problem as a runtime minimization problem, which is challenging to solve in practice due to a scene with different data centers. To address these challenges, we propose a model based on congestion-aware scheduling. In the model, we exploit SDN(Software-Defined Networking) to detect the data size of the data flow in advance from different data centers and then analyze the data characteristics, which predicts the flow that can generate network congestion in advance, so that we can draft two scheme for different flow. In addition, when we detect the network congestion, we choose a path with a greater bandwidth for the congestion flow. The approach can minimize network congestion, promote network utilization and improve system performance in a geo-distributed environment. As a highlight of this paper, we design and implement our proposed solution as a job scheduler based on Apache Spark, a modern data processing framework.
{"title":"Octopus: Based on Congestion-aware Scheduling on Geo-distributed Big Data Analytics Cluster","authors":"Haizhou Du, Keke Zhang, Zhenchen Yang","doi":"10.1109/ICSAI.2018.8599476","DOIUrl":"https://doi.org/10.1109/ICSAI.2018.8599476","url":null,"abstract":"In recent years, big data analytics frameworks spring up rapidly. Meanwhile, it has become routine for large volumes of data to be generated, stored, and processed across geographically distributed datac enters. Network congestion generated by data transfers between networks becomes a major bottleneck to the overall performance of the system in a geo-distributed environment. Many existing methods usually process network congestion after they occurs, which does not solve the problem fundamentally. In this paper, we focus on the problem of predicting and avoiding network congestion in advance in a geo-distributed environment on Apache Spark, in terms of their job completion times. We formulate this problem as a runtime minimization problem, which is challenging to solve in practice due to a scene with different data centers. To address these challenges, we propose a model based on congestion-aware scheduling. In the model, we exploit SDN(Software-Defined Networking) to detect the data size of the data flow in advance from different data centers and then analyze the data characteristics, which predicts the flow that can generate network congestion in advance, so that we can draft two scheme for different flow. In addition, when we detect the network congestion, we choose a path with a greater bandwidth for the congestion flow. The approach can minimize network congestion, promote network utilization and improve system performance in a geo-distributed environment. As a highlight of this paper, we design and implement our proposed solution as a job scheduler based on Apache Spark, a modern data processing framework.","PeriodicalId":375852,"journal":{"name":"2018 5th International Conference on Systems and Informatics (ICSAI)","volume":"31 3","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120860196","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 : 2018-11-01DOI: 10.1109/ICSAI.2018.8599366
Haizhou Du, Jingu Qian
Text classification is a fundamental problem in natural language processing. Recently, neural network models have been demonstrated to be capable of achieving remarkable performance in this domain. However, none of existing method can achieve excellent classification accuracy while concerning of computational cost. To solve this problem, we proposed hierarchical gated convolutional networks with multi-head attention which reduces computational cost through its two distinctive characteristics to save considerable model parameters. First, it has a hierarchical structure the same as the hierarchical structure of documents that has word-level and sentence-level, which not only benefits to classification performance but also reduces computational cost significantly by reusing parameters of the model in each sentence. Second, we apply gated convolutional network on both levels that enables our model achieved comparable performance to very deep networks with relatively shallow network depth. To further improve the performance of our model, multi-head attention mechanism is employed to differentiate more or less importance of words or sentences for better construction of document representation. Experiments conducted on the commonly used Yelp reviews datasets demonstrate that the proposed architecture obtains competitive performance against the state-of-the-art methods.
{"title":"Hierarchical Gated Convolutional Networks with Multi-Head Attention for Text Classification","authors":"Haizhou Du, Jingu Qian","doi":"10.1109/ICSAI.2018.8599366","DOIUrl":"https://doi.org/10.1109/ICSAI.2018.8599366","url":null,"abstract":"Text classification is a fundamental problem in natural language processing. Recently, neural network models have been demonstrated to be capable of achieving remarkable performance in this domain. However, none of existing method can achieve excellent classification accuracy while concerning of computational cost. To solve this problem, we proposed hierarchical gated convolutional networks with multi-head attention which reduces computational cost through its two distinctive characteristics to save considerable model parameters. First, it has a hierarchical structure the same as the hierarchical structure of documents that has word-level and sentence-level, which not only benefits to classification performance but also reduces computational cost significantly by reusing parameters of the model in each sentence. Second, we apply gated convolutional network on both levels that enables our model achieved comparable performance to very deep networks with relatively shallow network depth. To further improve the performance of our model, multi-head attention mechanism is employed to differentiate more or less importance of words or sentences for better construction of document representation. Experiments conducted on the commonly used Yelp reviews datasets demonstrate that the proposed architecture obtains competitive performance against the state-of-the-art methods.","PeriodicalId":375852,"journal":{"name":"2018 5th International Conference on Systems and Informatics (ICSAI)","volume":"85 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127987690","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：