Pub Date : 2002-11-18DOI: 10.1109/ICONIP.2002.1202142
A. Samarin, Y. Igarashi, I. B. Kulagina, S. Korogod
Neurons taken from animal embryo and grown on plane substrate with integrated multi-electrode array possess spontaneous bursts of action potentials, which become increasingly synchronous in the course of development. We studied such synchronization phenomena in a network of simulated neurons, which accounted for developmental changes in geometry and membrane properties of biological prototypes. Growing neuritic fields of individual neurons were partitioned into dendritic and axonal sectors of random size and orientation. Synapses were established between the cells with intersecting axonal and dendritic sectors and the synaptic weight was set in proportion to the intersection area. The network was imported into NEURON simulator. The membrane properties and the intracellular calcium dynamics of the cells were set similar to those described for the neocortical pyramidal neurons. Before being connected the cells generated asynchronous bursts. The synchronization developed in the domains of interconnected cells had the patterns apparently linked to the connectivity patterns.
{"title":"Activity synchronization in neural networks developing on planar substrates","authors":"A. Samarin, Y. Igarashi, I. B. Kulagina, S. Korogod","doi":"10.1109/ICONIP.2002.1202142","DOIUrl":"https://doi.org/10.1109/ICONIP.2002.1202142","url":null,"abstract":"Neurons taken from animal embryo and grown on plane substrate with integrated multi-electrode array possess spontaneous bursts of action potentials, which become increasingly synchronous in the course of development. We studied such synchronization phenomena in a network of simulated neurons, which accounted for developmental changes in geometry and membrane properties of biological prototypes. Growing neuritic fields of individual neurons were partitioned into dendritic and axonal sectors of random size and orientation. Synapses were established between the cells with intersecting axonal and dendritic sectors and the synaptic weight was set in proportion to the intersection area. The network was imported into NEURON simulator. The membrane properties and the intracellular calcium dynamics of the cells were set similar to those described for the neocortical pyramidal neurons. Before being connected the cells generated asynchronous bursts. The synchronization developed in the domains of interconnected cells had the patterns apparently linked to the connectivity patterns.","PeriodicalId":146553,"journal":{"name":"Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02.","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130688386","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 : 2002-11-18DOI: 10.1109/ICONIP.2002.1202195
M. Yoshida, H. Hayashi
Regulations of spontaneous rhythmic activity and preserved stimulus dependent patterns by spike timing dependent synaptic plasticity (STDP) were investigated in the hippocampal CA3 network model. With the presence of STDP, the rhythmic activity such as the theta rhythm might modify recurrent connections in the hippocampal CA3. In this study, we applied STDP to a hippocampal CA3 network model that causes spontaneous rhythmic activity. As a result, some local regions appeared in the network by themselves, from which the spontaneous rhythmic activity propagated toward surrounding area (source of propagation). We found that the frequency of the spontaneous rhythmic activity converged into one specific frequency depending on the shape of the STDP modification function. We also found that burst stimulation with sufficiently high frequency of bursts could produce a source of propagation at the stimulus site. The period of time for which the source of propagation was preserved after the termination of the stimulation depended on the shape of the STDP modification function.
{"title":"Regulation of spontaneous rhythmic activity and preserved stimulus dependent pattern by STDP in the hippocampal CA3 model","authors":"M. Yoshida, H. Hayashi","doi":"10.1109/ICONIP.2002.1202195","DOIUrl":"https://doi.org/10.1109/ICONIP.2002.1202195","url":null,"abstract":"Regulations of spontaneous rhythmic activity and preserved stimulus dependent patterns by spike timing dependent synaptic plasticity (STDP) were investigated in the hippocampal CA3 network model. With the presence of STDP, the rhythmic activity such as the theta rhythm might modify recurrent connections in the hippocampal CA3. In this study, we applied STDP to a hippocampal CA3 network model that causes spontaneous rhythmic activity. As a result, some local regions appeared in the network by themselves, from which the spontaneous rhythmic activity propagated toward surrounding area (source of propagation). We found that the frequency of the spontaneous rhythmic activity converged into one specific frequency depending on the shape of the STDP modification function. We also found that burst stimulation with sufficiently high frequency of bursts could produce a source of propagation at the stimulus site. The period of time for which the source of propagation was preserved after the termination of the stimulation depended on the shape of the STDP modification function.","PeriodicalId":146553,"journal":{"name":"Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02.","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132581596","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 : 2002-11-18DOI: 10.1109/ICONIP.2002.1198199
G. A. Khuwaja
Biometrics, which refers to identifying an individual based on his or her physiological or behavioral characteristics, has the capability to reliably distinguish between an authorized person and an imposter. This paper presents a high speed and low cost system for identification of fingerprints based on adaptive learning vector quantization neural network. The inkless images are acquired for this purpose using a digital still camera.
{"title":"Fingerprint identification with LVQ","authors":"G. A. Khuwaja","doi":"10.1109/ICONIP.2002.1198199","DOIUrl":"https://doi.org/10.1109/ICONIP.2002.1198199","url":null,"abstract":"Biometrics, which refers to identifying an individual based on his or her physiological or behavioral characteristics, has the capability to reliably distinguish between an authorized person and an imposter. This paper presents a high speed and low cost system for identification of fingerprints based on adaptive learning vector quantization neural network. The inkless images are acquired for this purpose using a digital still camera.","PeriodicalId":146553,"journal":{"name":"Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02.","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133158269","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 : 2002-11-18DOI: 10.1109/ICONIP.2002.1202799
S. Morokami, A. Suemitsu, M. Morita
In the delayed match-to-sample task, responses of inferior temporal neurons to adjacent stimuli in the sequence are correlated to each other when the monkey was trained repeatedly with the sequence of visual stimuli, although the monkey was not required to associate the stimuli with each other. This correlation, however, is not observed for a monkey with lesions of the rhinal cortex, which is not consistently explained by existing models of such correlated responses. In the present study, we construct a model consisting of two networks corresponding to area TE and the perirhinal cortex, and show that perirhinal plasticity may underlie the mechanism of implicit association learning.
{"title":"A model of implicit association learning based on plasticity in the perirhinal cortex","authors":"S. Morokami, A. Suemitsu, M. Morita","doi":"10.1109/ICONIP.2002.1202799","DOIUrl":"https://doi.org/10.1109/ICONIP.2002.1202799","url":null,"abstract":"In the delayed match-to-sample task, responses of inferior temporal neurons to adjacent stimuli in the sequence are correlated to each other when the monkey was trained repeatedly with the sequence of visual stimuli, although the monkey was not required to associate the stimuli with each other. This correlation, however, is not observed for a monkey with lesions of the rhinal cortex, which is not consistently explained by existing models of such correlated responses. In the present study, we construct a model consisting of two networks corresponding to area TE and the perirhinal cortex, and show that perirhinal plasticity may underlie the mechanism of implicit association learning.","PeriodicalId":146553,"journal":{"name":"Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02.","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133185387","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 : 2002-11-18DOI: 10.1109/ICONIP.2002.1202840
Zhiyong Liu, L. Xu
We propose a topological local principal component analysis (PCA) in help of Kohonen's self-organizing maps (SOM). The topological local PCA describes one cluster by one neuron such that it is capable of exploiting both the global topological structure and each local cluster structure. We also investigate a newly proposed self-organizing strategy that can enhance the learning speed, as well as an alternative Stiefel manifold based algorithm to ensure the orthonormality constraint of the local PCA.
{"title":"Topological local principal component analysis","authors":"Zhiyong Liu, L. Xu","doi":"10.1109/ICONIP.2002.1202840","DOIUrl":"https://doi.org/10.1109/ICONIP.2002.1202840","url":null,"abstract":"We propose a topological local principal component analysis (PCA) in help of Kohonen's self-organizing maps (SOM). The topological local PCA describes one cluster by one neuron such that it is capable of exploiting both the global topological structure and each local cluster structure. We also investigate a newly proposed self-organizing strategy that can enhance the learning speed, as well as an alternative Stiefel manifold based algorithm to ensure the orthonormality constraint of the local PCA.","PeriodicalId":146553,"journal":{"name":"Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02.","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128837497","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 : 2002-11-18DOI: 10.1109/ICONIP.2002.1202124
H. Urakubo, M. Watanabe
We simulate pairing backpropagating action potentials (APs) with excitatory postsynaptic potentials (EPSPs) in a hippocampal CA1 model neuron, and make clear the synaptic-site dependence of the supralinear amplification, which is considered as the origin of LTP in spike-timing dependent plasticity (STDP). Pairing APs with EPSPs at randomly determined synaptic sites reveals that the supralinear amplification needs modest APs with strong EPSPs, and the sites are restricted in apical middle dendrites and basal distal dendrites with small diameters. We also discuss the importance of this restriction in hippocampal information processing.
{"title":"Condition of supralinear amplification in pairing action potentials with EPSPS","authors":"H. Urakubo, M. Watanabe","doi":"10.1109/ICONIP.2002.1202124","DOIUrl":"https://doi.org/10.1109/ICONIP.2002.1202124","url":null,"abstract":"We simulate pairing backpropagating action potentials (APs) with excitatory postsynaptic potentials (EPSPs) in a hippocampal CA1 model neuron, and make clear the synaptic-site dependence of the supralinear amplification, which is considered as the origin of LTP in spike-timing dependent plasticity (STDP). Pairing APs with EPSPs at randomly determined synaptic sites reveals that the supralinear amplification needs modest APs with strong EPSPs, and the sites are restricted in apical middle dendrites and basal distal dendrites with small diameters. We also discuss the importance of this restriction in hippocampal information processing.","PeriodicalId":146553,"journal":{"name":"Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02.","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131676855","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 : 2002-11-18DOI: 10.1109/ICONIP.2002.1198162
T. Aaoyagi, H. Tokutaka, K. Fujimura, Y. Maniwa
We consider fast independent component analysis (FastICA), which is one of the independent component analysis algorithms. FastICA was proposed by Aapo Hyvarinen et al., (2001). It adopts the method of extracting the independent components one after another by the batch method using kurtosis. This method has fast convergence. The purpose of this research is to apply FastICA to the feature extraction of pulse waves of a human being, and to verify its effectiveness. The pulse waves contain a lot of information concerning the circulation of the blood from the heart to the various parts of the body. When blood flows from the heart and is transmitted to the tips as a wave motion, it is modified by physiological conditions such as the heart beat movement, the circulation of the blood flow, and changes in the state of the minor artery system, which leads to the distortion of the shape of the waves. The individual distortions have been evaluated and several trials have been performed to evaluate the health of a person. SOM is used to cluster the pulse waves and the features extracted from each cluster are considered.
{"title":"Application of FastICA to pulse wave","authors":"T. Aaoyagi, H. Tokutaka, K. Fujimura, Y. Maniwa","doi":"10.1109/ICONIP.2002.1198162","DOIUrl":"https://doi.org/10.1109/ICONIP.2002.1198162","url":null,"abstract":"We consider fast independent component analysis (FastICA), which is one of the independent component analysis algorithms. FastICA was proposed by Aapo Hyvarinen et al., (2001). It adopts the method of extracting the independent components one after another by the batch method using kurtosis. This method has fast convergence. The purpose of this research is to apply FastICA to the feature extraction of pulse waves of a human being, and to verify its effectiveness. The pulse waves contain a lot of information concerning the circulation of the blood from the heart to the various parts of the body. When blood flows from the heart and is transmitted to the tips as a wave motion, it is modified by physiological conditions such as the heart beat movement, the circulation of the blood flow, and changes in the state of the minor artery system, which leads to the distortion of the shape of the waves. The individual distortions have been evaluated and several trials have been performed to evaluate the health of a person. SOM is used to cluster the pulse waves and the features extracted from each cluster are considered.","PeriodicalId":146553,"journal":{"name":"Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02.","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115357527","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 : 2002-11-18DOI: 10.1109/ICONIP.2002.1199022
A.S. d'Avila Garcez, L. Lamb, D. Gabbay
Neural-Symbolic integration has become a very active research area in the last decade. In this paper, we present a new massively parallel model for modal logic. We do so by extending the language of Modal Prolog to allow modal operators in the head of the clauses. We then use an ensemble of C-IL/sup 2/p neural networks to encode the extended modal theory (and its relations), and show that the ensemble computes a fixpoint semantics of the extended theory. An immediate result of our approach is the ability to perform learning from examples efficiently using each network of the ensemble. Therefore, one can adapt the extended C-IL/sup 2/P system by training possible world representations.
{"title":"A connectionist inductive learning system for modal logic programming","authors":"A.S. d'Avila Garcez, L. Lamb, D. Gabbay","doi":"10.1109/ICONIP.2002.1199022","DOIUrl":"https://doi.org/10.1109/ICONIP.2002.1199022","url":null,"abstract":"Neural-Symbolic integration has become a very active research area in the last decade. In this paper, we present a new massively parallel model for modal logic. We do so by extending the language of Modal Prolog to allow modal operators in the head of the clauses. We then use an ensemble of C-IL/sup 2/p neural networks to encode the extended modal theory (and its relations), and show that the ensemble computes a fixpoint semantics of the extended theory. An immediate result of our approach is the ability to perform learning from examples efficiently using each network of the ensemble. Therefore, one can adapt the extended C-IL/sup 2/P system by training possible world representations.","PeriodicalId":146553,"journal":{"name":"Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02.","volume":"117 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115588650","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 : 2002-11-18DOI: 10.1109/ICONIP.2002.1198178
K. Hosoya, T. Ogawa, H. Kanada, K. Mori
The method to estimate the feature of the material by the impact sound was proposed ((M. Sakata and H. Ohnabe, 1994). To design the structure of composites taking into account the characteristic of the ceramics, a method was proposed to obtain the elastic moduli and the dumping ratio from the vibration of the material. To estimate their parameters, it is necessary to model the vibration precisely. In previous work, the vibration is analyzed by the fast Fourier transforms. On the other hand, the artificial neural network has been used to model the signal source, recently. The multilayer neural network adaptively models the signal source by error backpropagation. We propose a new neural network model for vibrational analysis of the material. We examined the model by the vibration waveform of actual ceramics composite. Also, the waveform at the high temperature is analyzed from the impact sound waveform of room temperature.
提出了用撞击声估计材料特性的方法(M. Sakata and H. Ohnabe, 1994)。为了设计考虑陶瓷特性的复合材料结构，提出了一种从材料振动中获得弹性模量和倾倒比的方法。为了估计它们的参数，必须精确地建立振动模型。在以前的工作中，用快速傅里叶变换来分析振动。另一方面，近年来人工神经网络已被用于信号源的建模。多层神经网络通过误差反向传播对信号源进行自适应建模。我们提出了一种新的用于材料振动分析的神经网络模型。用实际陶瓷复合材料的振动波形对模型进行了验证。并从常温下的冲击声波形出发，分析了高温下的冲击声波形。
{"title":"A neural network model for analyzing vibration waveform of impact sound","authors":"K. Hosoya, T. Ogawa, H. Kanada, K. Mori","doi":"10.1109/ICONIP.2002.1198178","DOIUrl":"https://doi.org/10.1109/ICONIP.2002.1198178","url":null,"abstract":"The method to estimate the feature of the material by the impact sound was proposed ((M. Sakata and H. Ohnabe, 1994). To design the structure of composites taking into account the characteristic of the ceramics, a method was proposed to obtain the elastic moduli and the dumping ratio from the vibration of the material. To estimate their parameters, it is necessary to model the vibration precisely. In previous work, the vibration is analyzed by the fast Fourier transforms. On the other hand, the artificial neural network has been used to model the signal source, recently. The multilayer neural network adaptively models the signal source by error backpropagation. We propose a new neural network model for vibrational analysis of the material. We examined the model by the vibration waveform of actual ceramics composite. Also, the waveform at the high temperature is analyzed from the impact sound waveform of room temperature.","PeriodicalId":146553,"journal":{"name":"Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02.","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115614057","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 : 2002-11-18DOI: 10.1109/ICONIP.2002.1202175
A. Kumar, V. Manmohan, M. Uday Shankar, M. Vishwanathan, V. Chakravarthy
Linking information processing and energy flows via thermodynamics, Landauer (1961) proposed that irreversible computational processes have an inevitable "thermodynamic cost". We explore the existence of such a link in case of a neural network model of associative memory. Our simulations with an electronic implementation of the Hopfield neural network showed that enhanced performance of the network could only be obtained by increased dissipation of energy as heat. Contrarily, efforts to minimize energy dissipation led to impaired performance.
{"title":"Link between energy and computation in a physical model of Hopfield network","authors":"A. Kumar, V. Manmohan, M. Uday Shankar, M. Vishwanathan, V. Chakravarthy","doi":"10.1109/ICONIP.2002.1202175","DOIUrl":"https://doi.org/10.1109/ICONIP.2002.1202175","url":null,"abstract":"Linking information processing and energy flows via thermodynamics, Landauer (1961) proposed that irreversible computational processes have an inevitable \"thermodynamic cost\". We explore the existence of such a link in case of a neural network model of associative memory. Our simulations with an electronic implementation of the Hopfield neural network showed that enhanced performance of the network could only be obtained by increased dissipation of energy as heat. Contrarily, efforts to minimize energy dissipation led to impaired performance.","PeriodicalId":146553,"journal":{"name":"Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02.","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114303446","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：