Pub Date : 2020-11-01DOI: 10.1109/CIS52066.2020.00028
Zhen Wang, Zhenting Zhang
Meta-learning methods accomplish rapid adaptation to a new task using few samples by first learning an internal representation that matches with similar tasks. In this paper, we focus on few-shot relation extraction. Previous works in few-shot relation extraction aim at predicting the relation for a pair of entities in a sentence by training with a few labeled examples in each relation. However, these algorithms implicitly require that the meta-training tasks be mutually-exclusive, such that no single model can solve all of the tasks at once, which hampers the generalization ability of these methods. To more effectively generalize to new relations, in this paper we address this challenge by designing a meta-regularization objective. We propose a novel Bayesian meta-learning approach to effectively learn the prototype vectors of relations via regularization on weights, and a graph neural network (GNN) is used to parameterize the initial prior of the prototype vectors on the global relation graph. Our approach substantially outperforms standard algorithms, and experiments demonstrate the effectiveness of our proposed approach.
{"title":"Graph-based Bayesian Meta Relation Extraction","authors":"Zhen Wang, Zhenting Zhang","doi":"10.1109/CIS52066.2020.00028","DOIUrl":"https://doi.org/10.1109/CIS52066.2020.00028","url":null,"abstract":"Meta-learning methods accomplish rapid adaptation to a new task using few samples by first learning an internal representation that matches with similar tasks. In this paper, we focus on few-shot relation extraction. Previous works in few-shot relation extraction aim at predicting the relation for a pair of entities in a sentence by training with a few labeled examples in each relation. However, these algorithms implicitly require that the meta-training tasks be mutually-exclusive, such that no single model can solve all of the tasks at once, which hampers the generalization ability of these methods. To more effectively generalize to new relations, in this paper we address this challenge by designing a meta-regularization objective. We propose a novel Bayesian meta-learning approach to effectively learn the prototype vectors of relations via regularization on weights, and a graph neural network (GNN) is used to parameterize the initial prior of the prototype vectors on the global relation graph. Our approach substantially outperforms standard algorithms, and experiments demonstrate the effectiveness of our proposed approach.","PeriodicalId":106959,"journal":{"name":"2020 16th International Conference on Computational Intelligence and Security (CIS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128567004","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/CIS52066.2020.00064
Erping Song, Hecheng Li, Cuo Wanma
The constrained multi-objective optimization problems (CMOPs) is widely used in real-world applications and always hard to handle especially when the objective number becomes more or the constraints are too stringent. In this manuscript, an improved differential evolution method (IDEM) is proposed based on CMOEA/D as well as newly designed mutation operators. Firstly, one mutation operator is presented to improve infeasible points, in which any infeasible point is taken to divide other points into three groups by using the constraint violation information, and based on the division, a potential better point can be found and utilized to improve other infeasible points by the mutation operation. Then the other mutation operator is provided by designing an objective sorting scheme as well as an individual selection method. These two mutation operators are alternately and self- adaptively adopted in evolution process. Finally, the proposed algorithm is executed on some recent benchmark functions and compared with four state-of-the-art EMO algorithms. The experimental results show that IDEM can efficiently solve the CMOPs.
{"title":"An Improved Differential Evolution for Constrained Multi-Objective Optimization Problems","authors":"Erping Song, Hecheng Li, Cuo Wanma","doi":"10.1109/CIS52066.2020.00064","DOIUrl":"https://doi.org/10.1109/CIS52066.2020.00064","url":null,"abstract":"The constrained multi-objective optimization problems (CMOPs) is widely used in real-world applications and always hard to handle especially when the objective number becomes more or the constraints are too stringent. In this manuscript, an improved differential evolution method (IDEM) is proposed based on CMOEA/D as well as newly designed mutation operators. Firstly, one mutation operator is presented to improve infeasible points, in which any infeasible point is taken to divide other points into three groups by using the constraint violation information, and based on the division, a potential better point can be found and utilized to improve other infeasible points by the mutation operation. Then the other mutation operator is provided by designing an objective sorting scheme as well as an individual selection method. These two mutation operators are alternately and self- adaptively adopted in evolution process. Finally, the proposed algorithm is executed on some recent benchmark functions and compared with four state-of-the-art EMO algorithms. The experimental results show that IDEM can efficiently solve the CMOPs.","PeriodicalId":106959,"journal":{"name":"2020 16th International Conference on Computational Intelligence and Security (CIS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122412252","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/CIS52066.2020.00038
Kaihao Lin, Junyan Chen, Aoge Chen, Hu Huang
With the continuous development of artificial intelligence technology, deep learning technology is used to process a large number of real-time traffic scene information helping the management of public transportation, and traffic flow statistics can reflect the real-time traffic conditions. The paper uses the EfficientDet target detection algorithm to detect and analyze the traffic video frame information and carry out statistics of vehicle and pedestrian flow at traffic intersections.The system can calculate the vehicle speed and perceive the degree of traffic congestion in real-time. It's convenient for the traffic department to increase the utilization rate of the road.
{"title":"Application of the EfficientDet Algorithm in Traffic Flow Statistics","authors":"Kaihao Lin, Junyan Chen, Aoge Chen, Hu Huang","doi":"10.1109/CIS52066.2020.00038","DOIUrl":"https://doi.org/10.1109/CIS52066.2020.00038","url":null,"abstract":"With the continuous development of artificial intelligence technology, deep learning technology is used to process a large number of real-time traffic scene information helping the management of public transportation, and traffic flow statistics can reflect the real-time traffic conditions. The paper uses the EfficientDet target detection algorithm to detect and analyze the traffic video frame information and carry out statistics of vehicle and pedestrian flow at traffic intersections.The system can calculate the vehicle speed and perceive the degree of traffic congestion in real-time. It's convenient for the traffic department to increase the utilization rate of the road.","PeriodicalId":106959,"journal":{"name":"2020 16th International Conference on Computational Intelligence and Security (CIS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127703460","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/CIS52066.2020.00066
Wenlong Cao, Fan Huang, Mengce Zheng, Honggang Hu
Field-programmable gate arrays (FPGAs) are widely used in many fields because of their low power consumption, easy design and good performance. For applications running on FPGAs, security is very important. A lot of researches have been done on the security issue of FPGA implementations, many attacks and countermeasures have been proposed. The dual complementary strategy is a countermeasure designed to thwart side channel attacks. In this paper, we perform Correlation Power Analysis (CPA) against dual complementary AES implemented on the SAKURA-G FPGA board. For dual complementary AES with constant Hamming Weight (HW) value, which is demonstrated to be robust against CPA based on HW model, we successfully recover the secret key using Hamming Distance (HD) and Switching Distance (SD) models with 2,000 power traces. For dual complementary AES with constant HD, 16,000 resp. 10,000 power traces are required to recover the key with HD resp. SD model.
{"title":"Attacking FPGA-based Dual Complementary AES Implementation Using HD and SD Models","authors":"Wenlong Cao, Fan Huang, Mengce Zheng, Honggang Hu","doi":"10.1109/CIS52066.2020.00066","DOIUrl":"https://doi.org/10.1109/CIS52066.2020.00066","url":null,"abstract":"Field-programmable gate arrays (FPGAs) are widely used in many fields because of their low power consumption, easy design and good performance. For applications running on FPGAs, security is very important. A lot of researches have been done on the security issue of FPGA implementations, many attacks and countermeasures have been proposed. The dual complementary strategy is a countermeasure designed to thwart side channel attacks. In this paper, we perform Correlation Power Analysis (CPA) against dual complementary AES implemented on the SAKURA-G FPGA board. For dual complementary AES with constant Hamming Weight (HW) value, which is demonstrated to be robust against CPA based on HW model, we successfully recover the secret key using Hamming Distance (HD) and Switching Distance (SD) models with 2,000 power traces. For dual complementary AES with constant HD, 16,000 resp. 10,000 power traces are required to recover the key with HD resp. SD model.","PeriodicalId":106959,"journal":{"name":"2020 16th International Conference on Computational Intelligence and Security (CIS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132842993","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/CIS52066.2020.00085
Cen Caichun, He Jie, Guo Hui
Aiming at the problem concerning low determination efficiency of sub-mode position caused by asymmetry in image segmentation with an algorithm in the asymmetric square NAM model image representation, with the aim to realize rapid image marking, the present study improves a peculiar rasterized array data structure, so as to make it satisfy geometric computation. Then, based on the neighbor search algorithm by employing SNAM, a connected domain marking algorithm based on SNAM is proposed. Through comparison of experimental results concerning handling of different binary images with the pixel scanning based connected domain marking algorithm, the connected domain marking algorithm based on quadtree scanning and the connected domain marking algorithm based on SNAM, it is proved that the connected domain marking algorithm based on asymmetric square NAM has high efficiency.
{"title":"Connected Domain Algorithm Based on Asymmetric Square NAM","authors":"Cen Caichun, He Jie, Guo Hui","doi":"10.1109/CIS52066.2020.00085","DOIUrl":"https://doi.org/10.1109/CIS52066.2020.00085","url":null,"abstract":"Aiming at the problem concerning low determination efficiency of sub-mode position caused by asymmetry in image segmentation with an algorithm in the asymmetric square NAM model image representation, with the aim to realize rapid image marking, the present study improves a peculiar rasterized array data structure, so as to make it satisfy geometric computation. Then, based on the neighbor search algorithm by employing SNAM, a connected domain marking algorithm based on SNAM is proposed. Through comparison of experimental results concerning handling of different binary images with the pixel scanning based connected domain marking algorithm, the connected domain marking algorithm based on quadtree scanning and the connected domain marking algorithm based on SNAM, it is proved that the connected domain marking algorithm based on asymmetric square NAM has high efficiency.","PeriodicalId":106959,"journal":{"name":"2020 16th International Conference on Computational Intelligence and Security (CIS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134319461","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/CIS52066.2020.00077
HuChunAn, WenHao
The principle of differential evolutionary algorithm is easy to understand, and it has the advantages of fast convergence, simple operation and good stability, which has been favored by many researchers. However, the differential evolution algorithm is easy to fall into the local optimum, and even cause the algorithm to stagnate, the low efficiency, and the unstable convergence speed of algorithm. This paper proposes an improved differential evolution (SAF-DE) algorithm, which uses the perturbation formula to perturb the individual values in the population to make individual more diversified. So as to achieve the purpose of improving the accuracy and convergence speed in the optimization process of the differential evolution algorithm. algorithm, the improved algorithm has higher convergence speed and accuracy on some standard functions.
{"title":"A differential evolution SAF-DE algorithm which jumps out of local optimal","authors":"HuChunAn, WenHao","doi":"10.1109/CIS52066.2020.00077","DOIUrl":"https://doi.org/10.1109/CIS52066.2020.00077","url":null,"abstract":"The principle of differential evolutionary algorithm is easy to understand, and it has the advantages of fast convergence, simple operation and good stability, which has been favored by many researchers. However, the differential evolution algorithm is easy to fall into the local optimum, and even cause the algorithm to stagnate, the low efficiency, and the unstable convergence speed of algorithm. This paper proposes an improved differential evolution (SAF-DE) algorithm, which uses the perturbation formula to perturb the individual values in the population to make individual more diversified. So as to achieve the purpose of improving the accuracy and convergence speed in the optimization process of the differential evolution algorithm. algorithm, the improved algorithm has higher convergence speed and accuracy on some standard functions.","PeriodicalId":106959,"journal":{"name":"2020 16th International Conference on Computational Intelligence and Security (CIS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133204279","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/CIS52066.2020.00047
Huihui Qian, Wensheng Chen, Binbin Pan, Bo Chen
Kernel-based non-negative matrix factorization (KNMF) can non-linearly extract non-negative features for image-data representation and classification. However, different kernel functions would lead to different performance. This means that selecting an appropriate kernel function plays an important role in KNMF algorithms. In this paper, we construct a novel Mercer kernel function, called cosine kernel function, which has the advantages of translation invariance and robustness to noise. Based on the self-constructed cosine kernel, we further propose a cosine kernel-based NMF (CKNMF) approach. The iterative formulas of CKNMF are deduced using the gradient descent method. We empirically validate that our CKNMF algorithm is convergent. Compared with some state of the art kernel-based algorithms, experimental results indicate that the proposed CKNMF algorithm achieves superior performance on face recognition.
{"title":"Kernel Non-Negative Matrix Factorization Using Self-Constructed Cosine Kernel","authors":"Huihui Qian, Wensheng Chen, Binbin Pan, Bo Chen","doi":"10.1109/CIS52066.2020.00047","DOIUrl":"https://doi.org/10.1109/CIS52066.2020.00047","url":null,"abstract":"Kernel-based non-negative matrix factorization (KNMF) can non-linearly extract non-negative features for image-data representation and classification. However, different kernel functions would lead to different performance. This means that selecting an appropriate kernel function plays an important role in KNMF algorithms. In this paper, we construct a novel Mercer kernel function, called cosine kernel function, which has the advantages of translation invariance and robustness to noise. Based on the self-constructed cosine kernel, we further propose a cosine kernel-based NMF (CKNMF) approach. The iterative formulas of CKNMF are deduced using the gradient descent method. We empirically validate that our CKNMF algorithm is convergent. Compared with some state of the art kernel-based algorithms, experimental results indicate that the proposed CKNMF algorithm achieves superior performance on face recognition.","PeriodicalId":106959,"journal":{"name":"2020 16th International Conference on Computational Intelligence and Security (CIS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133225216","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/CIS52066.2020.00031
Chuanzi Li, Jining Feng, Li Hu, Junhong Li, Haibin Ma
With the continuous development of deep learning technology, neural networks such as convolutional neural network (CNN) have shown good performance in many fields, such as image processing. Meanwhile, the relevant algorithm has made great progress. But the experiment results show that the deeper the network layers, the more the number of parameters that need to be trained in neural network, and the massive computing resources will be consumed to reconstruct and train the deep convolutional neural network (DCNN) model. These parameters often need to be trained in large dataset. But in many practical applications, the effective sample dataset that can be collected are usually small and lack of annotated samples. It is a pity that models that perform well on large datasets often have overfitting problems when applied to small datasets. And transfer learning can recognize and apply knowledge and skills learned in previous domains/tasks to novel domains/tasks. Combining deep convolutional neural network learning with transfer learning can make full use of existing models with good performance to solve problems in new fields, so has received considerable attentions due to its high research value and wide application prospect. This paper focuses on the combination of CNN and transfer learning, analyzes their characteristics, summarizes the relevant models, methods and applications, so as to promote their effective fusion in image classification.
{"title":"Review of Image Classification Method Based on Deep Transfer Learning","authors":"Chuanzi Li, Jining Feng, Li Hu, Junhong Li, Haibin Ma","doi":"10.1109/CIS52066.2020.00031","DOIUrl":"https://doi.org/10.1109/CIS52066.2020.00031","url":null,"abstract":"With the continuous development of deep learning technology, neural networks such as convolutional neural network (CNN) have shown good performance in many fields, such as image processing. Meanwhile, the relevant algorithm has made great progress. But the experiment results show that the deeper the network layers, the more the number of parameters that need to be trained in neural network, and the massive computing resources will be consumed to reconstruct and train the deep convolutional neural network (DCNN) model. These parameters often need to be trained in large dataset. But in many practical applications, the effective sample dataset that can be collected are usually small and lack of annotated samples. It is a pity that models that perform well on large datasets often have overfitting problems when applied to small datasets. And transfer learning can recognize and apply knowledge and skills learned in previous domains/tasks to novel domains/tasks. Combining deep convolutional neural network learning with transfer learning can make full use of existing models with good performance to solve problems in new fields, so has received considerable attentions due to its high research value and wide application prospect. This paper focuses on the combination of CNN and transfer learning, analyzes their characteristics, summarizes the relevant models, methods and applications, so as to promote their effective fusion in image classification.","PeriodicalId":106959,"journal":{"name":"2020 16th International Conference on Computational Intelligence and Security (CIS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131782748","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/CIS52066.2020.00059
Xiaona Lu, He Li
Intuitionistic fuzzy preference relations (IFPRs), as an effective tool for expressing pair-wise comparisons between alternatives, has a great advantage in group decision making (GDM). The manuscript first compares and analyzes the existing definitions of additive consistency for IFPRs and points out the shortcomings of the existing definitions of additive consistency for IFPRs. To fill these gaps, a new definition of additive consistency for IFPRs is proposed Meanwhile, a mathematical programming model is presented to check whether an IFPR is additively consistent or not. In addition, for an inconsistent IFPR, a mathematical programming model is put forward to improve the consistent level. Furthermore, based on the mathematical programming model, a new individual marking decision method with IFPR is proposed, and the merits of the proposed method is analyzed through a associated example.
{"title":"A novel additive consistency for intuitionistic fuzzy preference relations","authors":"Xiaona Lu, He Li","doi":"10.1109/CIS52066.2020.00059","DOIUrl":"https://doi.org/10.1109/CIS52066.2020.00059","url":null,"abstract":"Intuitionistic fuzzy preference relations (IFPRs), as an effective tool for expressing pair-wise comparisons between alternatives, has a great advantage in group decision making (GDM). The manuscript first compares and analyzes the existing definitions of additive consistency for IFPRs and points out the shortcomings of the existing definitions of additive consistency for IFPRs. To fill these gaps, a new definition of additive consistency for IFPRs is proposed Meanwhile, a mathematical programming model is presented to check whether an IFPR is additively consistent or not. In addition, for an inconsistent IFPR, a mathematical programming model is put forward to improve the consistent level. Furthermore, based on the mathematical programming model, a new individual marking decision method with IFPR is proposed, and the merits of the proposed method is analyzed through a associated example.","PeriodicalId":106959,"journal":{"name":"2020 16th International Conference on Computational Intelligence and Security (CIS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123029568","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/CIS52066.2020.00083
Tao Yu, Zhen Liu, Yuaning Liu, Huai-bin Wang, N. Adilov
In recent years, due to the increased frequency of cyber-attacks, the negative impacts of cyber-attacks on society have increased. Therefore, the research on cyber-security and prevention of cyber-attacks, including intrusion detection as an effective means of defense against cyber-attacks, is warranted. Both in the research and in the development of the systems for intrusion detection, the machine learning and deep learning methods are widely utilized, and the NSL-KDD dataset is frequently used in algorithm research and verification. In this paper, we propose a new two-stage dimensionality reduction (TSDR) feature selection method and verified by NSL-KDD dataset. The method reduces the dimensionality of the dataset and significantly improves the calculation efficiency. The KNN algorithm is used to verify that the new feature selection method improves the calculation efficiency. The accuracy rate is only slightly reduced when compared to the full feature calculation.
{"title":"A New Feature Selection Method for Intrusion Detection System Dataset – TSDR method","authors":"Tao Yu, Zhen Liu, Yuaning Liu, Huai-bin Wang, N. Adilov","doi":"10.1109/CIS52066.2020.00083","DOIUrl":"https://doi.org/10.1109/CIS52066.2020.00083","url":null,"abstract":"In recent years, due to the increased frequency of cyber-attacks, the negative impacts of cyber-attacks on society have increased. Therefore, the research on cyber-security and prevention of cyber-attacks, including intrusion detection as an effective means of defense against cyber-attacks, is warranted. Both in the research and in the development of the systems for intrusion detection, the machine learning and deep learning methods are widely utilized, and the NSL-KDD dataset is frequently used in algorithm research and verification. In this paper, we propose a new two-stage dimensionality reduction (TSDR) feature selection method and verified by NSL-KDD dataset. The method reduces the dimensionality of the dataset and significantly improves the calculation efficiency. The KNN algorithm is used to verify that the new feature selection method improves the calculation efficiency. The accuracy rate is only slightly reduced when compared to the full feature calculation.","PeriodicalId":106959,"journal":{"name":"2020 16th International Conference on Computational Intelligence and Security (CIS)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128795507","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}