Pub Date : 2022-01-01DOI: 10.1109/SSCI51031.2022.10022239
G. Prévost, Éric Jacopin, T. Cazenave, C. Guettier
{"title":"Planning for millions of NPCs in Real-Time","authors":"G. Prévost, Éric Jacopin, T. Cazenave, C. Guettier","doi":"10.1109/SSCI51031.2022.10022239","DOIUrl":"https://doi.org/10.1109/SSCI51031.2022.10022239","url":null,"abstract":"","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"52 1","pages":"330-336"},"PeriodicalIF":0.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75557723","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 : 2021-01-01DOI: 10.1109/SSCI50451.2021.9659884
G. Ferreira, P. A. Sesini, Luis Paulo Brasil de Souza, A. Kubrusly, H. V. Ayala
{"title":"Corrosion-like Defect Severity Estimation in Pipelines Using Convolutional Neural Networks","authors":"G. Ferreira, P. A. Sesini, Luis Paulo Brasil de Souza, A. Kubrusly, H. V. Ayala","doi":"10.1109/SSCI50451.2021.9659884","DOIUrl":"https://doi.org/10.1109/SSCI50451.2021.9659884","url":null,"abstract":"","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"26 1","pages":"1-7"},"PeriodicalIF":0.0,"publicationDate":"2021-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83272705","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 : 2021-01-01DOI: 10.1109/SSCI50451.2021.9660047
José Luis Fernández Pérez, L. M. Mariño, Roberto S. M. Barros
{"title":"Improving Diversity in Concept Drift Ensembles","authors":"José Luis Fernández Pérez, L. M. Mariño, Roberto S. M. Barros","doi":"10.1109/SSCI50451.2021.9660047","DOIUrl":"https://doi.org/10.1109/SSCI50451.2021.9660047","url":null,"abstract":"","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"60 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2021-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73977161","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 : 2021-01-01DOI: 10.1109/SSCI50451.2021.9660015
Murilo Oliveira Machado, E. Goldbarg, M. Goldbarg, Gustavo de Araujo Sabry, I. F. C. Fernandes, Thiago Soares Marques
{"title":"Heuristic Hybridization for CaRSP, a multilevel decision problem","authors":"Murilo Oliveira Machado, E. Goldbarg, M. Goldbarg, Gustavo de Araujo Sabry, I. F. C. Fernandes, Thiago Soares Marques","doi":"10.1109/SSCI50451.2021.9660015","DOIUrl":"https://doi.org/10.1109/SSCI50451.2021.9660015","url":null,"abstract":"","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"68 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2021-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89256352","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 : 2021-01-01DOI: 10.1109/SSCI50451.2021.9659940
Ericks da Silva Rodrigues, D. Martins, Fernando Buarque de Lima-Neto
{"title":"Self-Organizing Transformations for Automatic Feature Engineering","authors":"Ericks da Silva Rodrigues, D. Martins, Fernando Buarque de Lima-Neto","doi":"10.1109/SSCI50451.2021.9659940","DOIUrl":"https://doi.org/10.1109/SSCI50451.2021.9659940","url":null,"abstract":"","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"26 1","pages":"1-7"},"PeriodicalIF":0.0,"publicationDate":"2021-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77875491","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002703
Lei Lu, Y. Tan, D. Oetomo, I. Mareels, Erying Zhao, Shi An
Many techniques exist to train neural networks to approximate a complex systems, such as deep learning methods. It is well known that despite their robustness with respect to over-fitting, such trained models may be brittle as some fundamental principles of the systems are missing. For many engineering applications, the model class may be derived from first principles (or fundamental principles). In this paper, ideas from both methodologies are combined to arrive at a robust model that is interpretable from first principles, but goes beyond this by capturing structure from the available data. The paper presents two examples to illustrate the ideas. First a synthetic data set based on simulations is used. Next a well known data set from functional magnetic resonance imaging (fMRI) is used. In these two examples, a few representative neural networks are used in combination with model information coming from first principles. The preliminary results show that the framework is highly beneficial and yields excellent system identification fidelity.
{"title":"On Model-guided Neural Networks for System Identification","authors":"Lei Lu, Y. Tan, D. Oetomo, I. Mareels, Erying Zhao, Shi An","doi":"10.1109/SSCI44817.2019.9002703","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002703","url":null,"abstract":"Many techniques exist to train neural networks to approximate a complex systems, such as deep learning methods. It is well known that despite their robustness with respect to over-fitting, such trained models may be brittle as some fundamental principles of the systems are missing. For many engineering applications, the model class may be derived from first principles (or fundamental principles). In this paper, ideas from both methodologies are combined to arrive at a robust model that is interpretable from first principles, but goes beyond this by capturing structure from the available data. The paper presents two examples to illustrate the ideas. First a synthetic data set based on simulations is used. Next a well known data set from functional magnetic resonance imaging (fMRI) is used. In these two examples, a few representative neural networks are used in combination with model information coming from first principles. The preliminary results show that the framework is highly beneficial and yields excellent system identification fidelity.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"222 1","pages":"610-616"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75885855","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003060
Lu Luo, Shukai Duan, Lidan Wang
Convolutional neural networks (CNNs) have achieved the best performance in some fields. However, they still have some defects. CNNs need a lot of images for training; they will lose much information in the pooling layer, which reduces the spatial resolution. Facing such problems, Hinton et al. proposed a capsule network (CapsNet). Although the CapsNet has achieved the best accuracy on MNIST dataset, it has not performed well on Fashion-MNIST, Cifar-10 and other datasets. Naturally, we established an improved version of capsule network (R-CapsNet). Results have shown that when using R-CapsNet model, the loss gets decreased and the accuracy gets improved on FashionMNIST. In the meanwhile, the training parameters are reduced by nearly half. Specifically, it reduces by 4.5M. Comparisons show that our proposed model reports improved accuracy of around 0.56% over the existing state-of-the-art systems in literature. The test accuracy of R-CapsNet model is 1.32% higher than that of the original model. Furthermore, better results have been achieved on Cifar-10 with R-CapsNet model and it has easily increased by 10% compared to CapsNet.
{"title":"R-CapsNet: An Improvement of Capsule Network for More Complex Data","authors":"Lu Luo, Shukai Duan, Lidan Wang","doi":"10.1109/SSCI44817.2019.9003060","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003060","url":null,"abstract":"Convolutional neural networks (CNNs) have achieved the best performance in some fields. However, they still have some defects. CNNs need a lot of images for training; they will lose much information in the pooling layer, which reduces the spatial resolution. Facing such problems, Hinton et al. proposed a capsule network (CapsNet). Although the CapsNet has achieved the best accuracy on MNIST dataset, it has not performed well on Fashion-MNIST, Cifar-10 and other datasets. Naturally, we established an improved version of capsule network (R-CapsNet). Results have shown that when using R-CapsNet model, the loss gets decreased and the accuracy gets improved on FashionMNIST. In the meanwhile, the training parameters are reduced by nearly half. Specifically, it reduces by 4.5M. Comparisons show that our proposed model reports improved accuracy of around 0.56% over the existing state-of-the-art systems in literature. The test accuracy of R-CapsNet model is 1.32% higher than that of the original model. Furthermore, better results have been achieved on Cifar-10 with R-CapsNet model and it has easily increased by 10% compared to CapsNet.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"56 1","pages":"2124-2129"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74693264","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002765
Feng Zhang, Xinye Cai, Zhun Fan
A batched constrained decomposition with grids (BCDG) is proposed for expensive multiobjective optimization problems. In this algorithm, each objective function is approximated by a Gaussian process model and CDG-MOEA is used to optimize a candidate population. Finally, we use Hypervolume Indicator to select some better points from the candidate population for evaluation. In the process of CDG-MOEA optimizing candidate solutions and using Hypervolume Indicator to select candidate solutions for evaluation, we use Gaussian process lower confidence bound criteria to consider the uncertainty of Gaussian process prediction. Experimental study on some special test problems shows that BCDG can effectively solve some special expensive multiobjective optimization problems.
{"title":"A Batched Expensive Multiobjective Optimization Based on Constrained Decomposition with Grids","authors":"Feng Zhang, Xinye Cai, Zhun Fan","doi":"10.1109/SSCI44817.2019.9002765","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002765","url":null,"abstract":"A batched constrained decomposition with grids (BCDG) is proposed for expensive multiobjective optimization problems. In this algorithm, each objective function is approximated by a Gaussian process model and CDG-MOEA is used to optimize a candidate population. Finally, we use Hypervolume Indicator to select some better points from the candidate population for evaluation. In the process of CDG-MOEA optimizing candidate solutions and using Hypervolume Indicator to select candidate solutions for evaluation, we use Gaussian process lower confidence bound criteria to consider the uncertainty of Gaussian process prediction. Experimental study on some special test problems shows that BCDG can effectively solve some special expensive multiobjective optimization problems.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"2 1","pages":"2081-2087"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75731121","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}
X-ray safety inspection equipment is widely used in various public places for the detection of dangerous goods. At present, X-ray safety inspections mostly rely on manual inspections, so the detection efficiency is unsatisfactory. In the field of image detection technology, the deep learning based method has the advantages of low cost and simple configuration. In this paper ,we propose More Scales You Only Look Once version 3 (MS-YOLOv3) to detect and identify dangerous goods under X-ray.MS-YOLOv3 optimize the original You Only Look Once version 3 (YOLOv3) network structure by means of residual network and multi-scale fusion, improve the loss function and use the dangerous goods dataset under X-ray for training and testing. The experimental results show that the mAP of the optimized method is 7.08% higher than YOLOv3.
x射线安全检查设备广泛应用于各种公共场所，用于检测危险物品。目前，x射线安全检查大多依靠人工检查，检测效率差强人意。在图像检测技术领域，基于深度学习的方法具有成本低、配置简单等优点。在本文中，我们提出了更多的尺度你只看一次版本3 (MS-YOLOv3)来检测和识别x射线下的危险品。MS-YOLOv3通过残差网络和多尺度融合对原来的You Only Look Once version 3 (YOLOv3)网络结构进行优化，改进损失函数，并使用x射线下的危险品数据集进行训练和测试。实验结果表明，优化方法的mAP比YOLOv3提高了7.08%。
{"title":"Detection and Identification Method of Dangerous Goods under X-ray Based on Improved YOLOv3","authors":"Hong Zhang, Boyuan Xue, Qiang Zhi, Yiwen Fu, Lingfei Han, Qing Zhang, Chao Zhang","doi":"10.1109/SSCI44817.2019.9003131","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003131","url":null,"abstract":"X-ray safety inspection equipment is widely used in various public places for the detection of dangerous goods. At present, X-ray safety inspections mostly rely on manual inspections, so the detection efficiency is unsatisfactory. In the field of image detection technology, the deep learning based method has the advantages of low cost and simple configuration. In this paper ,we propose More Scales You Only Look Once version 3 (MS-YOLOv3) to detect and identify dangerous goods under X-ray.MS-YOLOv3 optimize the original You Only Look Once version 3 (YOLOv3) network structure by means of residual network and multi-scale fusion, improve the loss function and use the dangerous goods dataset under X-ray for training and testing. The experimental results show that the mAP of the optimized method is 7.08% higher than YOLOv3.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"51 1","pages":"2747-2752"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74052349","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 : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002680
Yichao Tang, Bo Wei, Xuewen Xia, Ling Gui
This paper presents a dynamic multi-swarm particle swarm optimization based on elite learning (DMS-PSO-EL) that consists of two kinds of sub-swarms to trade-off between exploitation and exploration capabilities. In DMS-PSO-EL, the whole population is divided into several DMS sub-swarms and one following sub-swarm on the basis of the fitness value rankings. In the evolution process, these DMS sub-swarms provide the exploration ability through dynamic regrouping strategy, while following sub-swarm enhances the exploitation ability by learning elite particles from DMS sub-swarms. Besides, randomly regrouping schedule regroups the entire population in each regrouping period aiming to avoid premature convergence and enhance inferior particles’ searching ability. Comparing DMSPSO-EL with other 8 peer algorithms on CEC2013 benchmark functions, the results suggest that DMS-PSO-EL demonstrates superior performance for solving different types of functions. Besides that, the massive experiments show the superiority of the proposed strategy used in DMS-PSO-EL.
{"title":"Dynamic Multi-swarm Particle Swarm Optimization Based on Elite Learning","authors":"Yichao Tang, Bo Wei, Xuewen Xia, Ling Gui","doi":"10.1109/SSCI44817.2019.9002680","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002680","url":null,"abstract":"This paper presents a dynamic multi-swarm particle swarm optimization based on elite learning (DMS-PSO-EL) that consists of two kinds of sub-swarms to trade-off between exploitation and exploration capabilities. In DMS-PSO-EL, the whole population is divided into several DMS sub-swarms and one following sub-swarm on the basis of the fitness value rankings. In the evolution process, these DMS sub-swarms provide the exploration ability through dynamic regrouping strategy, while following sub-swarm enhances the exploitation ability by learning elite particles from DMS sub-swarms. Besides, randomly regrouping schedule regroups the entire population in each regrouping period aiming to avoid premature convergence and enhance inferior particles’ searching ability. Comparing DMSPSO-EL with other 8 peer algorithms on CEC2013 benchmark functions, the results suggest that DMS-PSO-EL demonstrates superior performance for solving different types of functions. Besides that, the massive experiments show the superiority of the proposed strategy used in DMS-PSO-EL.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"128 1","pages":"2311-2318"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78599847","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：