Pub Date : 2021-11-25DOI: 10.1109/icicn52636.2021.9673847
Manal El Baz, T. Zaki, H. Douzi
The red color segmentation is widely used in image processing especially for traffic sign detection. The color segmentation under variety of light and weather conditions is a challenging task. These conditions make modifications on traffic sign colors. In this paper, we aim to improve the precision of a state-of-the-art method in RGB and HSV spaces. Although the efficiency of this state-of-the-art method, the red segmentation in RGB space has some false positive detections especially at night, and the red segmentation in HSV space detects the black pixels with red pixels as a region of interest. In order to improve the precision of this method, we propose in this paper to combine the two segmentations. We also propose changing the threshold of the red segmentation in HSV to remove non-red pixels and to extract red pixels with low saturation. In addition, we make some modifications on the threshold of the red segmentation in RGB space to improve its performance. The results obtained in experiments show that our method detects more accurately red color of traffic signs under variety of light and weather conditions compared to the state-of-the-art method.
{"title":"An Improved Method for Red Segmentation Based Traffic Sign Detection","authors":"Manal El Baz, T. Zaki, H. Douzi","doi":"10.1109/icicn52636.2021.9673847","DOIUrl":"https://doi.org/10.1109/icicn52636.2021.9673847","url":null,"abstract":"The red color segmentation is widely used in image processing especially for traffic sign detection. The color segmentation under variety of light and weather conditions is a challenging task. These conditions make modifications on traffic sign colors. In this paper, we aim to improve the precision of a state-of-the-art method in RGB and HSV spaces. Although the efficiency of this state-of-the-art method, the red segmentation in RGB space has some false positive detections especially at night, and the red segmentation in HSV space detects the black pixels with red pixels as a region of interest. In order to improve the precision of this method, we propose in this paper to combine the two segmentations. We also propose changing the threshold of the red segmentation in HSV to remove non-red pixels and to extract red pixels with low saturation. In addition, we make some modifications on the threshold of the red segmentation in RGB space to improve its performance. The results obtained in experiments show that our method detects more accurately red color of traffic signs under variety of light and weather conditions compared to the state-of-the-art method.","PeriodicalId":231379,"journal":{"name":"2021 IEEE 9th International Conference on Information, Communication and Networks (ICICN)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128254437","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 : 2021-11-25DOI: 10.1109/icicn52636.2021.9674018
Maria S. Papadopoulou, V. Rusyn, A. Boursianis, P. Sarigiannidis, Konstantinos E. Psannis, S. Goudos
It is common knowledge that nature exhibits nonlinear behavior, generally. As a result, non-linear phenomena play a vital role in the control systems in terms of engineering. That is the main reason for considering teaching nonlinear circuits in science and engineering undergraduate programs. In this paper, we present a well-known system that exhibits chaotic behavior. We quote the state equations and mathematical analysis of the Lorenz system. Afterward, we demonstrate simulation results to study and analyze the dynamic behavior of the overall system. The examples indicate a variety of ways in which chaotic behavior can arise in electronic circuits. Finally, we introduce an Arduino-based implementation of the Lorenz system. The comparison between the simulation and experimental results indicates the chaotic dynamics of the system.
{"title":"Diverse Implementations of the Lorenz System for Teaching Non-Linear Chaotic Circuits","authors":"Maria S. Papadopoulou, V. Rusyn, A. Boursianis, P. Sarigiannidis, Konstantinos E. Psannis, S. Goudos","doi":"10.1109/icicn52636.2021.9674018","DOIUrl":"https://doi.org/10.1109/icicn52636.2021.9674018","url":null,"abstract":"It is common knowledge that nature exhibits nonlinear behavior, generally. As a result, non-linear phenomena play a vital role in the control systems in terms of engineering. That is the main reason for considering teaching nonlinear circuits in science and engineering undergraduate programs. In this paper, we present a well-known system that exhibits chaotic behavior. We quote the state equations and mathematical analysis of the Lorenz system. Afterward, we demonstrate simulation results to study and analyze the dynamic behavior of the overall system. The examples indicate a variety of ways in which chaotic behavior can arise in electronic circuits. Finally, we introduce an Arduino-based implementation of the Lorenz system. The comparison between the simulation and experimental results indicates the chaotic dynamics of the system.","PeriodicalId":231379,"journal":{"name":"2021 IEEE 9th International Conference on Information, Communication and Networks (ICICN)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116632074","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 : 2021-11-25DOI: 10.1109/icicn52636.2021.9673923
Liu Zhen, Xin Shukai, Huang Xuguang, Liao Jianqing, Liang Guang
In view of the high cost of millimeter wave generated by the existing radio over fiber system, and the urgent need for high capacity and high bandwidth of mobile communication technology in the future, this paper proposes a microwave photonics system with 16quadrature amplitude modulation-orthogonal frequency division multiplexing (16QAM-OFDM) based on direct modulation and subcarrier multiplexing. The system adopts direct modulation technology combined with multiple subcarriers multiplexing, which includes 5G low frequency and millimeter-wave band. By transmitting multiple signals at the same time, the capacity of communication is expanded. In addition, OFDM is combined to further improve the spectral efficiency. Through simulation analysis, the results show that when the optical fiber transmission distance is 50km, the bit error rate (BER) of uplink and downlink is lower than $3.8 times 10^{-3}$. The constellation is clear and the error vector magnitude (EVM) value meets the standard that no more than 20% specified by the standard performance evaluation organization. The system provides a reference for ROF system that can be compatible with the two application scenarios of seamless wide-area coverage and high-capacity hot-spot.
{"title":"Microwave Photonics System with 16QAM-OFDM Based on Direct Modulation and Subcarrier Multiplexing","authors":"Liu Zhen, Xin Shukai, Huang Xuguang, Liao Jianqing, Liang Guang","doi":"10.1109/icicn52636.2021.9673923","DOIUrl":"https://doi.org/10.1109/icicn52636.2021.9673923","url":null,"abstract":"In view of the high cost of millimeter wave generated by the existing radio over fiber system, and the urgent need for high capacity and high bandwidth of mobile communication technology in the future, this paper proposes a microwave photonics system with 16quadrature amplitude modulation-orthogonal frequency division multiplexing (16QAM-OFDM) based on direct modulation and subcarrier multiplexing. The system adopts direct modulation technology combined with multiple subcarriers multiplexing, which includes 5G low frequency and millimeter-wave band. By transmitting multiple signals at the same time, the capacity of communication is expanded. In addition, OFDM is combined to further improve the spectral efficiency. Through simulation analysis, the results show that when the optical fiber transmission distance is 50km, the bit error rate (BER) of uplink and downlink is lower than $3.8 times 10^{-3}$. The constellation is clear and the error vector magnitude (EVM) value meets the standard that no more than 20% specified by the standard performance evaluation organization. The system provides a reference for ROF system that can be compatible with the two application scenarios of seamless wide-area coverage and high-capacity hot-spot.","PeriodicalId":231379,"journal":{"name":"2021 IEEE 9th International Conference on Information, Communication and Networks (ICICN)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122278633","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 : 2021-11-25DOI: 10.1109/icicn52636.2021.9673855
Tianming Xu, Song Song, Yahe Yang, Shuyu Xiao, Yejun Liu, Lei Guo
Free space optical (FSO) communication technology is a promising technology to provide higher bandwidth and faster transmission rate. However, FSO link is sensitive to various weather conditions and turbulence. Therefore, the relay-assisted FSO network has been paid much attention. Nevertheless, the existing relay selection schemes seldom consider the predicted link states and the schemes always have a time lag, which will reduce system performance. The paper proposes an intelligent relay selection mechanism for FSO communication. We introduce machine learning technology into channel and network traffic estimation to reduce the network outage probability. Experimental results demonstrate that the proposed scheme can efficiently reduce outage probability and increase network throughput.
{"title":"Intelligent Relay Selection with Predicted Link States for Relay-Assisted Free Space Optical Network","authors":"Tianming Xu, Song Song, Yahe Yang, Shuyu Xiao, Yejun Liu, Lei Guo","doi":"10.1109/icicn52636.2021.9673855","DOIUrl":"https://doi.org/10.1109/icicn52636.2021.9673855","url":null,"abstract":"Free space optical (FSO) communication technology is a promising technology to provide higher bandwidth and faster transmission rate. However, FSO link is sensitive to various weather conditions and turbulence. Therefore, the relay-assisted FSO network has been paid much attention. Nevertheless, the existing relay selection schemes seldom consider the predicted link states and the schemes always have a time lag, which will reduce system performance. The paper proposes an intelligent relay selection mechanism for FSO communication. We introduce machine learning technology into channel and network traffic estimation to reduce the network outage probability. Experimental results demonstrate that the proposed scheme can efficiently reduce outage probability and increase network throughput.","PeriodicalId":231379,"journal":{"name":"2021 IEEE 9th International Conference on Information, Communication and Networks (ICICN)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132105096","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 : 2021-11-25DOI: 10.1109/icicn52636.2021.9674023
Feng Shimin, Fu Tianhui, Yang Yongkang, Xiu Menglei, Wang Longfei, Wang Shiyu
The towed antenna enables the underwater vehicle to communicate across the sea while maintaining its invisibility. In this paper, a Ku-band towed antenna is proposed based on the large loop helix structure. The impedance characteristics of large loop helical antenna are deduced theoretically. Using FEKO electromagnetic simulation software, a layered model of air-sea-towed cable was established to simulate the electrical characteristics of the large loop spiral antenna. According to the simulation results, the length and the number of turns of the large loop spiral antenna are optimized. The results show that the large-loop helical towed antenna can effectively improve the Ku-band signal reception capability of the underwater vehicle in the underwater safe depth range.
{"title":"Design and Optimization of Ku-band Large-Loop-Spiral Towed Antenna for the Unmanned Underwater Vehicle","authors":"Feng Shimin, Fu Tianhui, Yang Yongkang, Xiu Menglei, Wang Longfei, Wang Shiyu","doi":"10.1109/icicn52636.2021.9674023","DOIUrl":"https://doi.org/10.1109/icicn52636.2021.9674023","url":null,"abstract":"The towed antenna enables the underwater vehicle to communicate across the sea while maintaining its invisibility. In this paper, a Ku-band towed antenna is proposed based on the large loop helix structure. The impedance characteristics of large loop helical antenna are deduced theoretically. Using FEKO electromagnetic simulation software, a layered model of air-sea-towed cable was established to simulate the electrical characteristics of the large loop spiral antenna. According to the simulation results, the length and the number of turns of the large loop spiral antenna are optimized. The results show that the large-loop helical towed antenna can effectively improve the Ku-band signal reception capability of the underwater vehicle in the underwater safe depth range.","PeriodicalId":231379,"journal":{"name":"2021 IEEE 9th International Conference on Information, Communication and Networks (ICICN)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132298273","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 : 2021-11-25DOI: 10.1109/icicn52636.2021.9673965
Kainan Ma, Tao Li, Yibo Yin, Sitao Zhang, Ming Liu
To accelerate the convergence of the soft-decision decoder and improve the reading performance of the NAND flash memory, a multi-level reading method is usually adopted to sense precise log-likelihood ratios (LLR). However, multi-level readings interfere with the threshold voltage in the cells, increasing the probability of bit errors and accelerating cells wear. To solve this problem, this paper proposed an LLR soft-sensing method based on a broad learning system (BLS) to replace multi-level reading, which improves accuracy of sensing from 97.1% to 97.3% by using the raw bit error rate (RBER) as one of the inputs. The output classification probabilities of the network model are used to calculate the LLR. Compared with the multilayer perceptron model, the adoption of BLS hugely decreases the computation amount of network training from 113 epochs to 5, making on-chip retraining feasible. The proposed BLS-based LLR soft-sensing method will be of great application potential in the intelligent error control of non-volatile memories.
{"title":"A Method of Soft-Sensing Log-Likelihood Ratios Based on Broad Learning System for NAND Flash Memories","authors":"Kainan Ma, Tao Li, Yibo Yin, Sitao Zhang, Ming Liu","doi":"10.1109/icicn52636.2021.9673965","DOIUrl":"https://doi.org/10.1109/icicn52636.2021.9673965","url":null,"abstract":"To accelerate the convergence of the soft-decision decoder and improve the reading performance of the NAND flash memory, a multi-level reading method is usually adopted to sense precise log-likelihood ratios (LLR). However, multi-level readings interfere with the threshold voltage in the cells, increasing the probability of bit errors and accelerating cells wear. To solve this problem, this paper proposed an LLR soft-sensing method based on a broad learning system (BLS) to replace multi-level reading, which improves accuracy of sensing from 97.1% to 97.3% by using the raw bit error rate (RBER) as one of the inputs. The output classification probabilities of the network model are used to calculate the LLR. Compared with the multilayer perceptron model, the adoption of BLS hugely decreases the computation amount of network training from 113 epochs to 5, making on-chip retraining feasible. The proposed BLS-based LLR soft-sensing method will be of great application potential in the intelligent error control of non-volatile memories.","PeriodicalId":231379,"journal":{"name":"2021 IEEE 9th International Conference on Information, Communication and Networks (ICICN)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131840719","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 : 2021-11-25DOI: 10.1109/icicn52636.2021.9673840
Mengyin Liu, Ziyang Mei, Gongli Xiao, Hongyan Yang
In this paper, we propose a defect coupled photonic crystal fiber (PCF) biosensor detection method. A double-layer annular pore array with different diameters is designed, which is characterized by six periodically arranged holes (d1=0.7um) in the inner layer of the core and five symmetrically arranged holes (d2=0.8um) in the outer layer. By removing two air holes to form defects, the evanescent field leakage to the pore defects can be effectively enhanced. By using the finite element method (FEM), the sensing performance can be greatly improved by optimizing the geometric structure parameters, metal thickness and adding graphene. When the refractive index of the measured medium changes from 1.33 to 1.37, the sensitivity of the sensor can be improved to 3000nm/RIU.
{"title":"Graphene-Photonic Crystal Fiber Biodetection Based on Surface Plasma Resonance Effect and Defect Coupling","authors":"Mengyin Liu, Ziyang Mei, Gongli Xiao, Hongyan Yang","doi":"10.1109/icicn52636.2021.9673840","DOIUrl":"https://doi.org/10.1109/icicn52636.2021.9673840","url":null,"abstract":"In this paper, we propose a defect coupled photonic crystal fiber (PCF) biosensor detection method. A double-layer annular pore array with different diameters is designed, which is characterized by six periodically arranged holes (d1=0.7um) in the inner layer of the core and five symmetrically arranged holes (d2=0.8um) in the outer layer. By removing two air holes to form defects, the evanescent field leakage to the pore defects can be effectively enhanced. By using the finite element method (FEM), the sensing performance can be greatly improved by optimizing the geometric structure parameters, metal thickness and adding graphene. When the refractive index of the measured medium changes from 1.33 to 1.37, the sensitivity of the sensor can be improved to 3000nm/RIU.","PeriodicalId":231379,"journal":{"name":"2021 IEEE 9th International Conference on Information, Communication and Networks (ICICN)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129378056","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 : 2021-11-25DOI: 10.1109/icicn52636.2021.9673934
Sana Mazhar, Shawwal Rashed, Saleem Zubair
Testing is the significant fragment of SDLC where programming is appropriately tried and changes are proposed. Performance testing is considered to manage the affectability, dependability, and extensibility of a plan under a particular outstanding task at hand. It is utilized owing to the increasing approval of the web service, web services are widely used today and there is little literature on the performance of web services. Web tenders are harder to trial than conventional tenders, particularly execution assessments, e.g., sudden burdens, reaction times, and so on. Testing a web application is nothing more than finding errors in its content, functions, usability, navigation, performance, capabilities and security. This article presents the concepts, objectives, goals and apparatuses accessible for execution testing of web applications. Different execution testing instruments are accessible. In this paper, we have attempted to discover the components that improve the presentation of a web application.
{"title":"Comparative Analysis of Software Performance on Web based Application","authors":"Sana Mazhar, Shawwal Rashed, Saleem Zubair","doi":"10.1109/icicn52636.2021.9673934","DOIUrl":"https://doi.org/10.1109/icicn52636.2021.9673934","url":null,"abstract":"Testing is the significant fragment of SDLC where programming is appropriately tried and changes are proposed. Performance testing is considered to manage the affectability, dependability, and extensibility of a plan under a particular outstanding task at hand. It is utilized owing to the increasing approval of the web service, web services are widely used today and there is little literature on the performance of web services. Web tenders are harder to trial than conventional tenders, particularly execution assessments, e.g., sudden burdens, reaction times, and so on. Testing a web application is nothing more than finding errors in its content, functions, usability, navigation, performance, capabilities and security. This article presents the concepts, objectives, goals and apparatuses accessible for execution testing of web applications. Different execution testing instruments are accessible. In this paper, we have attempted to discover the components that improve the presentation of a web application.","PeriodicalId":231379,"journal":{"name":"2021 IEEE 9th International Conference on Information, Communication and Networks (ICICN)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116092174","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 : 2021-11-25DOI: 10.1109/icicn52636.2021.9673897
F. Ouellette, Jin Huang, Shucheng Liu, Jianfeng Li
We propose and demonstrate a referencing scheme for dual-wavelength differential detection of tilted fiber Bragg grating sensors that compensates for drift in the laser diode wavelength and dynamic chirp, and results in 40 fm rms stability, and a refractive index change limit of detection of $3 times 10 ^{-6}$.
我们提出并演示了一种用于倾斜光纤光栅传感器双波长差分检测的参考方案,该方案补偿了激光二极管波长的漂移和动态啁啾,从而获得了40 fm rms的稳定性,并且检测的折射率变化极限为3 × 10 ^{-6}$。
{"title":"Measurement of Tilted Fiber Bragg Grating Sensors with Femtometer Resolution Using Dual-Wavelength Differential Detection","authors":"F. Ouellette, Jin Huang, Shucheng Liu, Jianfeng Li","doi":"10.1109/icicn52636.2021.9673897","DOIUrl":"https://doi.org/10.1109/icicn52636.2021.9673897","url":null,"abstract":"We propose and demonstrate a referencing scheme for dual-wavelength differential detection of tilted fiber Bragg grating sensors that compensates for drift in the laser diode wavelength and dynamic chirp, and results in 40 fm rms stability, and a refractive index change limit of detection of $3 times 10 ^{-6}$.","PeriodicalId":231379,"journal":{"name":"2021 IEEE 9th International Conference on Information, Communication and Networks (ICICN)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116645513","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 : 2021-11-25DOI: 10.1109/icicn52636.2021.9673867
Kun Yang, Haipeng Gao, Y. Yang, Ke Qin
Graph Neural Networks (GNNs) have recently been shown to be quite effective in modeling graph-structured data. Recent methods such as RGCN and SACN, have achieved the most advanced results in knowledge graph completion. However, previous efforts are mostly restricted to aggregating information given by neighboring nodes only, ignoring the information given by neighboring edges. This paper proposes a novel Entities and Relations Aware Graph Convolutional Network (ERA-GCN), with an encoder-decoder framework which jointly embeds both entities and relations in a multi-relation graph. In the encoder end, ERA-GCN uses a weighted graph convolutional network to capture both graph structure and neighborhood information. In the decoder end, we utilize Conv-TransE to retain the translational property between entity and relation embedding, leading to better link prediction performance. We evaluate our proposed method on standard FB15k-237 and WNISRR datasets, and achieve about 11% relative improvement compared to current state-of-the-art ConvE in terms of HITS@l, HITS@3 and HITS@10.
{"title":"Entities and Relations Aware Graph Convolutional Network for Knowledge Base Completion","authors":"Kun Yang, Haipeng Gao, Y. Yang, Ke Qin","doi":"10.1109/icicn52636.2021.9673867","DOIUrl":"https://doi.org/10.1109/icicn52636.2021.9673867","url":null,"abstract":"Graph Neural Networks (GNNs) have recently been shown to be quite effective in modeling graph-structured data. Recent methods such as RGCN and SACN, have achieved the most advanced results in knowledge graph completion. However, previous efforts are mostly restricted to aggregating information given by neighboring nodes only, ignoring the information given by neighboring edges. This paper proposes a novel Entities and Relations Aware Graph Convolutional Network (ERA-GCN), with an encoder-decoder framework which jointly embeds both entities and relations in a multi-relation graph. In the encoder end, ERA-GCN uses a weighted graph convolutional network to capture both graph structure and neighborhood information. In the decoder end, we utilize Conv-TransE to retain the translational property between entity and relation embedding, leading to better link prediction performance. We evaluate our proposed method on standard FB15k-237 and WNISRR datasets, and achieve about 11% relative improvement compared to current state-of-the-art ConvE in terms of HITS@l, HITS@3 and HITS@10.","PeriodicalId":231379,"journal":{"name":"2021 IEEE 9th International Conference on Information, Communication and Networks (ICICN)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121703050","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}