Pub Date : 2011-10-03DOI: 10.1109/IJCNN.2011.6033271
S. Montigny, Richard Labib
The quantron is a new artificial neuron model, able to solve nonlinear classification problems, for which an efficient learning algorithm has yet to be developed. Using surrogate potentials, constraints on some parameters and an infinite number of potentials, we obtain analytical expressions involving ceiling functions for the activation function of the quantron. We then show how to retrieve the parameters of a neuron from the images it produced.
{"title":"Learning algorithms for a specific configuration of the quantron","authors":"S. Montigny, Richard Labib","doi":"10.1109/IJCNN.2011.6033271","DOIUrl":"https://doi.org/10.1109/IJCNN.2011.6033271","url":null,"abstract":"The quantron is a new artificial neuron model, able to solve nonlinear classification problems, for which an efficient learning algorithm has yet to be developed. Using surrogate potentials, constraints on some parameters and an infinite number of potentials, we obtain analytical expressions involving ceiling functions for the activation function of the quantron. We then show how to retrieve the parameters of a neuron from the images it produced.","PeriodicalId":415833,"journal":{"name":"The 2011 International Joint Conference on Neural Networks","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132943847","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 : 2011-10-03DOI: 10.1109/IJCNN.2011.6033450
Riccardo Taormina, L. Mesin, Fiammetta Orione, E. Pasero
The issue of air quality is now a major concern for many citizens worldwide. Local air quality forecasting can be made on the basis of meteorological variables and air pollutants concentration time series. We propose an adaptive filter technique based on an artificial neural network (ANN) to make 24-hours maximal daily ozone-concentrations forecasts.
{"title":"Forecasting tropospheric ozone concentrations with adaptive neural networks","authors":"Riccardo Taormina, L. Mesin, Fiammetta Orione, E. Pasero","doi":"10.1109/IJCNN.2011.6033450","DOIUrl":"https://doi.org/10.1109/IJCNN.2011.6033450","url":null,"abstract":"The issue of air quality is now a major concern for many citizens worldwide. Local air quality forecasting can be made on the basis of meteorological variables and air pollutants concentration time series. We propose an adaptive filter technique based on an artificial neural network (ANN) to make 24-hours maximal daily ozone-concentrations forecasts.","PeriodicalId":415833,"journal":{"name":"The 2011 International Joint Conference on Neural Networks","volume":"144 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132985930","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 : 2011-10-03DOI: 10.1109/IJCNN.2011.6033247
Yuuji Ichisugi
We propose a computational model of recognition of the cerebral cortex, based on an approximate belief revision algorithm. The algorithm calculates the MPE (most probable explanation) of Bayesian networks with a linear-sum CPT (conditional probability table) model. Although the proposed algorithm is simple enough to be implemented by a fixed circuit, results of the performance evaluation show that this algorithm does not have bad approximation accuracy. The mean convergence time is not sensitive to the number of nodes if the depth the network is constant. The computation amount is linear to the number of nodes if the number of edges per node is constant. The proposed algorithm can be used as a part of a learning algorithm for a kind of sparse-coding, which reproduces orientation selectivity of the primary visual area. The circuit that executes the algorithm shows better correspondence to the anatomical structure of the cerebral cortex, namely its six-layer and columnar features, than the approximate belief propagation algorithm that has been proposed before. These results suggest that the proposed algorithm is a promising starting point for the model of the recognition mechanism of the cerebral cortex.
{"title":"Recognition model of cerebral cortex based on approximate belief revision algorithm","authors":"Yuuji Ichisugi","doi":"10.1109/IJCNN.2011.6033247","DOIUrl":"https://doi.org/10.1109/IJCNN.2011.6033247","url":null,"abstract":"We propose a computational model of recognition of the cerebral cortex, based on an approximate belief revision algorithm. The algorithm calculates the MPE (most probable explanation) of Bayesian networks with a linear-sum CPT (conditional probability table) model. Although the proposed algorithm is simple enough to be implemented by a fixed circuit, results of the performance evaluation show that this algorithm does not have bad approximation accuracy. The mean convergence time is not sensitive to the number of nodes if the depth the network is constant. The computation amount is linear to the number of nodes if the number of edges per node is constant. The proposed algorithm can be used as a part of a learning algorithm for a kind of sparse-coding, which reproduces orientation selectivity of the primary visual area. The circuit that executes the algorithm shows better correspondence to the anatomical structure of the cerebral cortex, namely its six-layer and columnar features, than the approximate belief propagation algorithm that has been proposed before. These results suggest that the proposed algorithm is a promising starting point for the model of the recognition mechanism of the cerebral cortex.","PeriodicalId":415833,"journal":{"name":"The 2011 International Joint Conference on Neural Networks","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132612651","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 : 2011-10-03DOI: 10.1109/IJCNN.2011.6033577
Sepideh Seifzadeh, Mohammad Rostami, A. Ghodsi, F. Karray
A challenging problem in smoothing spline regression is determining a value for the smoothing parameter. The parameter establishes the tradeoff between the closeness of the data, versus the smoothness of the regression function. This paper proposes a new method of finding the optimum smoothness value based on Stein's Unbiased Risk Estimator (SURE). This approach employs Newton's method to solve for the optimal value directly, while minimizing the true error of the regression. Experimental results demonstrate the effectiveness of this method, particularly for small datasets.
{"title":"Parameter selection for smoothing splines using Stein's Unbiased Risk Estimator","authors":"Sepideh Seifzadeh, Mohammad Rostami, A. Ghodsi, F. Karray","doi":"10.1109/IJCNN.2011.6033577","DOIUrl":"https://doi.org/10.1109/IJCNN.2011.6033577","url":null,"abstract":"A challenging problem in smoothing spline regression is determining a value for the smoothing parameter. The parameter establishes the tradeoff between the closeness of the data, versus the smoothness of the regression function. This paper proposes a new method of finding the optimum smoothness value based on Stein's Unbiased Risk Estimator (SURE). This approach employs Newton's method to solve for the optimal value directly, while minimizing the true error of the regression. Experimental results demonstrate the effectiveness of this method, particularly for small datasets.","PeriodicalId":415833,"journal":{"name":"The 2011 International Joint Conference on Neural Networks","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126552893","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 : 2011-10-03DOI: 10.1109/IJCNN.2011.6033318
Reghunadhan Rajesh, K. Rajeev, K. Suchithra, L. V. Prabhu, Vignesh Gopakumar, N. Ragesh
This paper makes use of Coherence Vector of Oriented Gradients (CVOG) for traffic sign recognition. Experiments are conducted on German Traffic Sign benchmark dataset. The results on traffic sign recognition using CVOG features with neural network classifier is promising. The results based on the combination of other features gave better recognition rates.
{"title":"Coherence vector of Oriented Gradients for traffic sign recognition using Neural Networks","authors":"Reghunadhan Rajesh, K. Rajeev, K. Suchithra, L. V. Prabhu, Vignesh Gopakumar, N. Ragesh","doi":"10.1109/IJCNN.2011.6033318","DOIUrl":"https://doi.org/10.1109/IJCNN.2011.6033318","url":null,"abstract":"This paper makes use of Coherence Vector of Oriented Gradients (CVOG) for traffic sign recognition. Experiments are conducted on German Traffic Sign benchmark dataset. The results on traffic sign recognition using CVOG features with neural network classifier is promising. The results based on the combination of other features gave better recognition rates.","PeriodicalId":415833,"journal":{"name":"The 2011 International Joint Conference on Neural Networks","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131246833","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 : 2011-10-03DOI: 10.1109/IJCNN.2011.6033556
R. Green, Lingfeng Wang, Mansoor Alam
Monte Carlo Simulation (MCS) is a very powerful and flexible tool when used for sampling states during the probabilistic reliability assessment of power systems. Despite the advantages of MCS, the method begins to falter when applied to large and more complex systems of higher dimensions. In these cases it is often the process of classifying states that consumes the majority of computational time and resources. This is especially true in power systems reliability evaluation where the main method of classification is typically an Optimal Power Flow (OPF) formulation in the form of a linear program (LP). Previous works have improved the computational time required for classification by using Neural Networks (NN) of varying types in place of the OPF. A method of classification that is lighter weight and often more computationally efficient than NNs is the Support Vector Machine (SVM). This work couples SVM with the MCS algorithm in order to improve the computational time of classification and overall reliability evaluation. The method is further extended through the use of a multi-core architecture in order to further decrease computational time. These formulations are tested using the IEEE Reliability Test Systems (IEEE-RTS79 and IEEE-RTS96). Significant improvements in computational time are demonstrated while a high level of accuracy is maintained.
{"title":"Composite power system reliability evaluation using support vector machines on a multicore platform","authors":"R. Green, Lingfeng Wang, Mansoor Alam","doi":"10.1109/IJCNN.2011.6033556","DOIUrl":"https://doi.org/10.1109/IJCNN.2011.6033556","url":null,"abstract":"Monte Carlo Simulation (MCS) is a very powerful and flexible tool when used for sampling states during the probabilistic reliability assessment of power systems. Despite the advantages of MCS, the method begins to falter when applied to large and more complex systems of higher dimensions. In these cases it is often the process of classifying states that consumes the majority of computational time and resources. This is especially true in power systems reliability evaluation where the main method of classification is typically an Optimal Power Flow (OPF) formulation in the form of a linear program (LP). Previous works have improved the computational time required for classification by using Neural Networks (NN) of varying types in place of the OPF. A method of classification that is lighter weight and often more computationally efficient than NNs is the Support Vector Machine (SVM). This work couples SVM with the MCS algorithm in order to improve the computational time of classification and overall reliability evaluation. The method is further extended through the use of a multi-core architecture in order to further decrease computational time. These formulations are tested using the IEEE Reliability Test Systems (IEEE-RTS79 and IEEE-RTS96). Significant improvements in computational time are demonstrated while a high level of accuracy is maintained.","PeriodicalId":415833,"journal":{"name":"The 2011 International Joint Conference on Neural Networks","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131244653","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}
Word burstiness phenomenon, which means that if a word occurs once in a document it is likely to occur repeatedly, has interested the text analysis field recently. Dirichlet Compound Multinomial Latent Dirichlet Allocation (DCMLDA) introduces this word burstiness mechanism into Latent Dirichlet Allocation (LDA). However, in DCMLDA, there is no restriction on the word burstiness intensity of each topic. Consequently, as shown in this paper, the burstiness intensities of words in major topics will become extremely low and the topics' ability to represent different semantic meanings will be impaired. In order to get topics that represent semantic meanings of documents well, we introduce constraints on topics' word burstiness intensities. Experiments demonstrate that DCMLDA with constrained word burstiness intensities achieves better performance than the original one without constraints. Besides, these additional constraints help to reveal the relationship between two key properties inherited from DCM and LDA respectively. These two properties have a great influence on the combined model's performance and their relationship revealed by this paper is an important guidance for further study of topic models.
{"title":"Topic model with constrainted word burstiness intensities","authors":"Shaoze Lei, Jianwen Zhang, Shifeng Weng, Changshui Zhang","doi":"10.1109/IJCNN.2011.6033201","DOIUrl":"https://doi.org/10.1109/IJCNN.2011.6033201","url":null,"abstract":"Word burstiness phenomenon, which means that if a word occurs once in a document it is likely to occur repeatedly, has interested the text analysis field recently. Dirichlet Compound Multinomial Latent Dirichlet Allocation (DCMLDA) introduces this word burstiness mechanism into Latent Dirichlet Allocation (LDA). However, in DCMLDA, there is no restriction on the word burstiness intensity of each topic. Consequently, as shown in this paper, the burstiness intensities of words in major topics will become extremely low and the topics' ability to represent different semantic meanings will be impaired. In order to get topics that represent semantic meanings of documents well, we introduce constraints on topics' word burstiness intensities. Experiments demonstrate that DCMLDA with constrained word burstiness intensities achieves better performance than the original one without constraints. Besides, these additional constraints help to reveal the relationship between two key properties inherited from DCM and LDA respectively. These two properties have a great influence on the combined model's performance and their relationship revealed by this paper is an important guidance for further study of topic models.","PeriodicalId":415833,"journal":{"name":"The 2011 International Joint Conference on Neural Networks","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132816638","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 : 2011-10-03DOI: 10.1109/IJCNN.2011.6033295
Takashi Matsubara, H. Torikai
A novel spiking neuron model whose nonlinear dynamics is described by an asynchronous cellular automaton is presented. The model can be implemented by a simple digital sequential logic circuit but can exhibit various neuron-like bifurcations and responses. Using the Poincaré mapping technique, it is clarified that the model can reproduce major bifurcation mechanisms of excitabilities and spikings of biological and model neurons. It is also clarified that the model can reproduce major excitatory responses of the neurons.
{"title":"A novel asynchronous digital spiking neuron model and its various neuron-like bifurcations and responses","authors":"Takashi Matsubara, H. Torikai","doi":"10.1109/IJCNN.2011.6033295","DOIUrl":"https://doi.org/10.1109/IJCNN.2011.6033295","url":null,"abstract":"A novel spiking neuron model whose nonlinear dynamics is described by an asynchronous cellular automaton is presented. The model can be implemented by a simple digital sequential logic circuit but can exhibit various neuron-like bifurcations and responses. Using the Poincaré mapping technique, it is clarified that the model can reproduce major bifurcation mechanisms of excitabilities and spikings of biological and model neurons. It is also clarified that the model can reproduce major excitatory responses of the neurons.","PeriodicalId":415833,"journal":{"name":"The 2011 International Joint Conference on Neural Networks","volume":"124 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132829328","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 : 2011-10-03DOI: 10.1109/IJCNN.2011.6033225
Fengchen Liu, Yongsheng Ding, Weixun Gao
Bio-networks have the natural advantages of autonomy, scalability, and adaptability which are challenges for computer networks, especially P2P networks. We present a bio-inspired dynamic balanced tree structure network (called bio-block) based dynamic network. Every bio-block is a unique bio-entities collection with emergent service. This network has two parts, non-Service part (bio-entity is unit node) and in-Service part (bio-block is unit node). Useful bio-entities are dynamically transferring between these two part to keep the balance, and improve resources usage. This network inherits the balanced structure and O(nlogN) search steps with total N resources and n resources service request. It also eliminates redundancies by taking advantage of strong adaptability of bio-network which are composed of bio-entities. Any node in this balanced tree structured network can join and leave dynamically. Intensive experimental results show that the state of this network is converged when service distribution is stable. Moreover, theoretical results support an efficient search operation.
{"title":"Bio-inspired balanced tree structure dynamic network","authors":"Fengchen Liu, Yongsheng Ding, Weixun Gao","doi":"10.1109/IJCNN.2011.6033225","DOIUrl":"https://doi.org/10.1109/IJCNN.2011.6033225","url":null,"abstract":"Bio-networks have the natural advantages of autonomy, scalability, and adaptability which are challenges for computer networks, especially P2P networks. We present a bio-inspired dynamic balanced tree structure network (called bio-block) based dynamic network. Every bio-block is a unique bio-entities collection with emergent service. This network has two parts, non-Service part (bio-entity is unit node) and in-Service part (bio-block is unit node). Useful bio-entities are dynamically transferring between these two part to keep the balance, and improve resources usage. This network inherits the balanced structure and O(nlogN) search steps with total N resources and n resources service request. It also eliminates redundancies by taking advantage of strong adaptability of bio-network which are composed of bio-entities. Any node in this balanced tree structured network can join and leave dynamically. Intensive experimental results show that the state of this network is converged when service distribution is stable. Moreover, theoretical results support an efficient search operation.","PeriodicalId":415833,"journal":{"name":"The 2011 International Joint Conference on Neural Networks","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132721832","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 : 2011-10-03DOI: 10.1109/IJCNN.2011.6033531
C. S. Pereira, George D. C. Cavalcanti
This paper presents an innovative instance selection method, called Instance Selection Algorithm based on a Ranking Procedure (ISAR), which is based on a ranking criterion. The ranking procedure aims to order the instances in the data set; better the instance higher the score associate to it. With the purpose of eliminating irrelevant instances, ISAR also uses a coverage strategy. Each instance delimits a hypersphere centered in it. The radius of each hypersphere is used as a normalization factor in the classification rule; bigger the radius smaller the distance. After a comparative study using real-world databases, the ISAR algorithm reached promising generalization performance and impressive reduction rates when compared with state of the art methods.
{"title":"Instance selection algorithm based on a Ranking Procedure","authors":"C. S. Pereira, George D. C. Cavalcanti","doi":"10.1109/IJCNN.2011.6033531","DOIUrl":"https://doi.org/10.1109/IJCNN.2011.6033531","url":null,"abstract":"This paper presents an innovative instance selection method, called Instance Selection Algorithm based on a Ranking Procedure (ISAR), which is based on a ranking criterion. The ranking procedure aims to order the instances in the data set; better the instance higher the score associate to it. With the purpose of eliminating irrelevant instances, ISAR also uses a coverage strategy. Each instance delimits a hypersphere centered in it. The radius of each hypersphere is used as a normalization factor in the classification rule; bigger the radius smaller the distance. After a comparative study using real-world databases, the ISAR algorithm reached promising generalization performance and impressive reduction rates when compared with state of the art methods.","PeriodicalId":415833,"journal":{"name":"The 2011 International Joint Conference on Neural Networks","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132733996","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：