Pub Date : 2013-09-08DOI: 10.1109/BRICS-CCI-CBIC.2013.71
Daniel S. F. Alves, D. O. Cardoso, Hugo C. C. Carneiro, F. França, P. Lima
This work investigates the effect of different data structures on the performance and accuracy of VG-RAM-based classifiers. This weightless neural model is based on RAM nodes having very large address input, what suggests the use of special data structures in order to deal with space and time computational costs. Four different data structures are explored, including the classical one used in recent VG-RAM related literature, resulting in a novel and accurate yet fast setup.
{"title":"An Empirical Study of the Influence of Data Structures on the Performance of VG-RAM Classifiers","authors":"Daniel S. F. Alves, D. O. Cardoso, Hugo C. C. Carneiro, F. França, P. Lima","doi":"10.1109/BRICS-CCI-CBIC.2013.71","DOIUrl":"https://doi.org/10.1109/BRICS-CCI-CBIC.2013.71","url":null,"abstract":"This work investigates the effect of different data structures on the performance and accuracy of VG-RAM-based classifiers. This weightless neural model is based on RAM nodes having very large address input, what suggests the use of special data structures in order to deal with space and time computational costs. Four different data structures are explored, including the classical one used in recent VG-RAM related literature, resulting in a novel and accurate yet fast setup.","PeriodicalId":306195,"journal":{"name":"2013 BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129368518","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 : 2013-09-08DOI: 10.1109/BRICS-CCI-CBIC.2013.94
Gabriela I. L. Alves, D. A. Silva, Emeson J. S. Pereira, T. Ferreira
Artificial Neural Networks (ANN) have been widely used in time series forecasting problem. However, a more promising approach is the combination of ANN with other intelligent techniques, such as genetic algorithms, evolutionary strategies, etc, where these evolutionary algorithms have the objective of train and adjust all parameter of the ANN. In the evolutionary process is necessary define a fitness function to guide the evolve process. Thus, for a set of possibles fitness function, how to determine the function more efficient? This paper aims to select the efficient fitness functions, through the use of Data Envelopment Analysis. This tool determines the relative efficiency of each unit under review, comparing it with each other and considering the relationship between inputs and outputs. Two different times series were used to benchmark the set of twenty fitness functions. The preliminary results show the proposed method is a promising approach for efficient selecting the fitness function.
{"title":"Data Envelopment Analysis for Selection of the Fitness Function in Evolutionary Algorithms Applied to Time Series Forecasting Problem","authors":"Gabriela I. L. Alves, D. A. Silva, Emeson J. S. Pereira, T. Ferreira","doi":"10.1109/BRICS-CCI-CBIC.2013.94","DOIUrl":"https://doi.org/10.1109/BRICS-CCI-CBIC.2013.94","url":null,"abstract":"Artificial Neural Networks (ANN) have been widely used in time series forecasting problem. However, a more promising approach is the combination of ANN with other intelligent techniques, such as genetic algorithms, evolutionary strategies, etc, where these evolutionary algorithms have the objective of train and adjust all parameter of the ANN. In the evolutionary process is necessary define a fitness function to guide the evolve process. Thus, for a set of possibles fitness function, how to determine the function more efficient? This paper aims to select the efficient fitness functions, through the use of Data Envelopment Analysis. This tool determines the relative efficiency of each unit under review, comparing it with each other and considering the relationship between inputs and outputs. Two different times series were used to benchmark the set of twenty fitness functions. The preliminary results show the proposed method is a promising approach for efficient selecting the fitness function.","PeriodicalId":306195,"journal":{"name":"2013 BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129579484","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 : 2013-09-08DOI: 10.1109/BRICS-CCI-CBIC.2013.95
P. Rougemont, Filipe Braida do Carmo, Marden Braga Pasinato, Carlos E. Mello, Geraldo Zimbrão
This work proposes a new methodology for the Group Recommendation problem. In this approach we choose the Most Representative User (MRU) as the group medoid in a user space projection, and then generate the recommendation list based on his preferences. We evaluate our proposal by using the well-known dataset Movie lens. We have taken two different measures so as to evaluate the group recommender strategies. The obtained results seem promising and our strategy has shown an empirical robustness compared with the baselines in the literature.
{"title":"Group Recommender Systems: Exploring Underlying Information of the User Space","authors":"P. Rougemont, Filipe Braida do Carmo, Marden Braga Pasinato, Carlos E. Mello, Geraldo Zimbrão","doi":"10.1109/BRICS-CCI-CBIC.2013.95","DOIUrl":"https://doi.org/10.1109/BRICS-CCI-CBIC.2013.95","url":null,"abstract":"This work proposes a new methodology for the Group Recommendation problem. In this approach we choose the Most Representative User (MRU) as the group medoid in a user space projection, and then generate the recommendation list based on his preferences. We evaluate our proposal by using the well-known dataset Movie lens. We have taken two different measures so as to evaluate the group recommender strategies. The obtained results seem promising and our strategy has shown an empirical robustness compared with the baselines in the literature.","PeriodicalId":306195,"journal":{"name":"2013 BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130933257","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 : 2013-09-08DOI: 10.1109/BRICS-CCI-CBIC.2013.46
Amauri H. Souza Junior, F. Corona, Y. Miché, A. Lendasse, G. Barreto
The Minimal Learning Machine (MLM) has been recently proposed as a novel supervised learning method for regression problems aiming at reconstructing the mapping between input and output distance matrices. Estimation of the response is then achieved from the geometrical configuration of the output points. Thanks to its comprehensive formulation, the MLM is inherently capable of dealing with nonlinear problems and multidimensional output spaces. In this paper, we introduce an extension of the MLM to classification tasks, thus providing a unified framework for multiresponse regression and classification problems. On the basis of our experiments, the MLM achieves results that are comparable to many de facto standard methods for classification with the advantage of offering a computationally lighter alternative to such approaches.
{"title":"Extending the Minimal Learning Machine for Pattern Classification","authors":"Amauri H. Souza Junior, F. Corona, Y. Miché, A. Lendasse, G. Barreto","doi":"10.1109/BRICS-CCI-CBIC.2013.46","DOIUrl":"https://doi.org/10.1109/BRICS-CCI-CBIC.2013.46","url":null,"abstract":"The Minimal Learning Machine (MLM) has been recently proposed as a novel supervised learning method for regression problems aiming at reconstructing the mapping between input and output distance matrices. Estimation of the response is then achieved from the geometrical configuration of the output points. Thanks to its comprehensive formulation, the MLM is inherently capable of dealing with nonlinear problems and multidimensional output spaces. In this paper, we introduce an extension of the MLM to classification tasks, thus providing a unified framework for multiresponse regression and classification problems. On the basis of our experiments, the MLM achieves results that are comparable to many de facto standard methods for classification with the advantage of offering a computationally lighter alternative to such approaches.","PeriodicalId":306195,"journal":{"name":"2013 BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130197856","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 : 2013-09-08DOI: 10.1109/BRICS-CCI-CBIC.2013.25
Thiago M. De Almeida, Reginaldo B. Nunes, Helder R. de O. Rocha, M. Segatto, Jair A. L. Silva
The employment of genetic algorithms in parameters optimization of direct-detection optical orthogonal frequency division multiplexing (DDO-OFDM) systems in short-range links is reported. Experimental transmission of a 3.56 Gb/s (4-QAM subcarrier mapping) optimized DDO-OFDM system in optical back-to-back (B2B) configuration and through 20 and 40 km of uncompensated standard single-mode fiber (SSMF) was achieved.
{"title":"Performance Optimization of DDO-OFDM Systems through Genetic Algorithms","authors":"Thiago M. De Almeida, Reginaldo B. Nunes, Helder R. de O. Rocha, M. Segatto, Jair A. L. Silva","doi":"10.1109/BRICS-CCI-CBIC.2013.25","DOIUrl":"https://doi.org/10.1109/BRICS-CCI-CBIC.2013.25","url":null,"abstract":"The employment of genetic algorithms in parameters optimization of direct-detection optical orthogonal frequency division multiplexing (DDO-OFDM) systems in short-range links is reported. Experimental transmission of a 3.56 Gb/s (4-QAM subcarrier mapping) optimized DDO-OFDM system in optical back-to-back (B2B) configuration and through 20 and 40 km of uncompensated standard single-mode fiber (SSMF) was achieved.","PeriodicalId":306195,"journal":{"name":"2013 BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125775525","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 : 2013-09-08DOI: 10.1109/BRICS-CCI-CBIC.2013.38
Diego F. P. De Souza, Hugo C. C. Carneiro, F. França, P. Lima
This paper presents some strategies used for creating intelligent players of rock-paper-scissors using WiSARD weightless neural networks and results obtained therewith. These strategies included: (i) a new approach for encoding of the input data, (ii) three new training algorithms that allow the reclassification of the input patterns over time, (iii) a method for dealing with incomplete information in the input array, and (iv) a bluffing strategy. Experiments show that, in a tournament of intelligent agents, WiSARD-based agents were ranked among the 200 best players, one of them achieving 9th place for about three weeks.
{"title":"Rock-Paper-Scissors WiSARD","authors":"Diego F. P. De Souza, Hugo C. C. Carneiro, F. França, P. Lima","doi":"10.1109/BRICS-CCI-CBIC.2013.38","DOIUrl":"https://doi.org/10.1109/BRICS-CCI-CBIC.2013.38","url":null,"abstract":"This paper presents some strategies used for creating intelligent players of rock-paper-scissors using WiSARD weightless neural networks and results obtained therewith. These strategies included: (i) a new approach for encoding of the input data, (ii) three new training algorithms that allow the reclassification of the input patterns over time, (iii) a method for dealing with incomplete information in the input array, and (iv) a bluffing strategy. Experiments show that, in a tournament of intelligent agents, WiSARD-based agents were ranked among the 200 best players, one of them achieving 9th place for about three weeks.","PeriodicalId":306195,"journal":{"name":"2013 BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131531974","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 : 2013-09-08DOI: 10.1109/BRICS-CCI-CBIC.2013.98
Wanessa da Silva, M. Habermann, Elcio Hideiti Shiguemori, Leidiane do Livramento Andrade, Ruy Morgado de Castro
This work presents a methodology for pattern classification from multispectral images acquired by the HSS airborne sensor. In order to achieve this purpose, a conjunction of Artificial Neural Network and Principal Components Analysis has been used. The results indicate that this approach can be alternatively employed in multispectral images to separate materials with specific characteristics based on their reflectance properties.
{"title":"Multispectral Image Classification Using Multilayer Perceptron and Principal Components Analysis","authors":"Wanessa da Silva, M. Habermann, Elcio Hideiti Shiguemori, Leidiane do Livramento Andrade, Ruy Morgado de Castro","doi":"10.1109/BRICS-CCI-CBIC.2013.98","DOIUrl":"https://doi.org/10.1109/BRICS-CCI-CBIC.2013.98","url":null,"abstract":"This work presents a methodology for pattern classification from multispectral images acquired by the HSS airborne sensor. In order to achieve this purpose, a conjunction of Artificial Neural Network and Principal Components Analysis has been used. The results indicate that this approach can be alternatively employed in multispectral images to separate materials with specific characteristics based on their reflectance properties.","PeriodicalId":306195,"journal":{"name":"2013 BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132478129","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 : 2013-09-08DOI: 10.1109/BRICS-CCI-CBIC.2013.50
David Iclanzan, Camelia Chira
The complex regimes of operation situated between ordered and chaotic behavior are hypothesized to give rise to computational capabilities. Lacking an universal blueprint for the emergence of complexity, a costly search is typically used to find the configurations of distributed artificial systems that can facilitate global computation. In this paper, we address the tedious task of searching for complex cellular automata rules able to lead to a certain global behavior based on local interactions. The discovery of rules exhibiting a high degree of global self-organization is of major importance in the study and understanding of complex systems. A classical heuristic search guided only by a coarse approximation of the ability of a rule to perform in certain conditions will generally not reach beyond an ordered regime of operation. To overcome this limitation, in this paper we incorporate a promising heuristic that rewards increased dynamics with regard to cell state changes in a multiobjective, parallel evolutionary framework. The scope of the multiobjective formulation is to balance the search between ordered and chaotic regimes in order to facilitate the discovery of rules exhibiting complex behaviors. Experimental results confirm that the combined approach represents an efficient way for supporting the emergence of complexity as in all runs we were able to find cellular automata exhibiting a high degree of global self-organization.
{"title":"A Parallel Multiobjective Approach to Evolving Cellular Automata Rules by Cell State Change Dynamics","authors":"David Iclanzan, Camelia Chira","doi":"10.1109/BRICS-CCI-CBIC.2013.50","DOIUrl":"https://doi.org/10.1109/BRICS-CCI-CBIC.2013.50","url":null,"abstract":"The complex regimes of operation situated between ordered and chaotic behavior are hypothesized to give rise to computational capabilities. Lacking an universal blueprint for the emergence of complexity, a costly search is typically used to find the configurations of distributed artificial systems that can facilitate global computation. In this paper, we address the tedious task of searching for complex cellular automata rules able to lead to a certain global behavior based on local interactions. The discovery of rules exhibiting a high degree of global self-organization is of major importance in the study and understanding of complex systems. A classical heuristic search guided only by a coarse approximation of the ability of a rule to perform in certain conditions will generally not reach beyond an ordered regime of operation. To overcome this limitation, in this paper we incorporate a promising heuristic that rewards increased dynamics with regard to cell state changes in a multiobjective, parallel evolutionary framework. The scope of the multiobjective formulation is to balance the search between ordered and chaotic regimes in order to facilitate the discovery of rules exhibiting complex behaviors. Experimental results confirm that the combined approach represents an efficient way for supporting the emergence of complexity as in all runs we were able to find cellular automata exhibiting a high degree of global self-organization.","PeriodicalId":306195,"journal":{"name":"2013 BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132346008","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 : 2013-09-08DOI: 10.1109/BRICS-CCI-CBIC.2013.30
A. Engelbrecht
Particle swarm optimization (PSO) is an iterative algorithm, where particle positions and best positions are updated per iteration. The order in which particle positions and best positions are updated is referred to in this paper as an iteration strategy. Two main iteration strategies exist for PSO, namely synchronous updates and asynchronous updates. A number of studies have discussed the advantages and disadvantages of these iteration strategies. Most of these studies indicated that asynchronous updates are better than synchronous updates with respect to accuracy of the solutions obtained and the speed at which swarms converge. This study provides evidence from an extensive empirical analysis that current opinions that asynchronous updates result in faster convergence and more accurate results are not true.
{"title":"Particle Swarm Optimization: Iteration Strategies Revisited","authors":"A. Engelbrecht","doi":"10.1109/BRICS-CCI-CBIC.2013.30","DOIUrl":"https://doi.org/10.1109/BRICS-CCI-CBIC.2013.30","url":null,"abstract":"Particle swarm optimization (PSO) is an iterative algorithm, where particle positions and best positions are updated per iteration. The order in which particle positions and best positions are updated is referred to in this paper as an iteration strategy. Two main iteration strategies exist for PSO, namely synchronous updates and asynchronous updates. A number of studies have discussed the advantages and disadvantages of these iteration strategies. Most of these studies indicated that asynchronous updates are better than synchronous updates with respect to accuracy of the solutions obtained and the speed at which swarms converge. This study provides evidence from an extensive empirical analysis that current opinions that asynchronous updates result in faster convergence and more accurate results are not true.","PeriodicalId":306195,"journal":{"name":"2013 BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121298972","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 : 2013-09-08DOI: 10.1109/BRICS-CCI-CBIC.2013.12
A. Averkin, V. Albu, S. Ulyanov, I. Povidalo
In this article a number of neural networks based on self organizing maps, that can be successfully used for dynamic object identification, is described. The structure and algorithms of learning and operation of such SOM-based neural networks are described in details, also some experimental results is given.
{"title":"Dynamic Object Identification with SOM-Based Neural Networks","authors":"A. Averkin, V. Albu, S. Ulyanov, I. Povidalo","doi":"10.1109/BRICS-CCI-CBIC.2013.12","DOIUrl":"https://doi.org/10.1109/BRICS-CCI-CBIC.2013.12","url":null,"abstract":"In this article a number of neural networks based on self organizing maps, that can be successfully used for dynamic object identification, is described. The structure and algorithms of learning and operation of such SOM-based neural networks are described in details, also some experimental results is given.","PeriodicalId":306195,"journal":{"name":"2013 BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124983554","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：