The Experts below are selected from a list of 116565 Experts worldwide ranked by ideXlab platform
Lei Deng - One of the best experts on this subject based on the ideXlab platform.
-
Direct Training for spiking neural networks faster larger better
National Conference on Artificial Intelligence, 2019Co-Authors: Lei Deng, Jun Zhu, Yuan Xie, Luping ShiAbstract:Spiking neural networks (SNNs) that enables energy efficient implementation on emerging neuromorphic hardware are gaining more attention. Yet now, SNNs have not shown competitive performance compared with artificial neural networks (ANNs), due to the lack of effective learning algorithms and efficient programming frameworks. We address this issue from two aspects: (1) We propose a neuron normalization technique to adjust the neural selectivity and develop a Direct learning algorithm for deep SNNs. (2) Via narrowing the rate coding window and converting the leaky integrate-and-fire (LIF) model into an explicitly iterative version, we present a Pytorch-based implementation method towards the Training of large-scale SNNs. In this way, we are able to train deep SNNs with tens of times speedup. As a result, we achieve significantly better accuracy than the reported works on neuromorphic datasets (N-MNIST and DVSCIFAR10), and comparable accuracy as existing ANNs and pre-trained SNNs on non-spiking datasets (CIFAR10). To our best knowledge, this is the first work that demonstrates Direct Training of deep SNNs with high performance on CIFAR10, and the efficient implementation provides a new way to explore the potential of SNNs.
-
spatio temporal backpropagation for Training high performance spiking neural networks
Frontiers in Neuroscience, 2018Co-Authors: Lei Deng, Jun Zhu, Luping ShiAbstract:Spiking neural networks (SNNs) are promising in ascertaining brain-like behaviors since spikes are capable of encoding spatio-temporal information. Recent schemes, e.g. pre-Training from artificial neural networks (ANNs) or Direct Training based on backpropagation (BP), make the high-performance supervised Training of SNNs possible. However, these methods primarily fasten more attention on its spatial domain information, and the dynamics in temporal domain are attached less significance. Consequently, this might lead to the performance bottleneck, and scores of Training techniques shall be additionally required. Another underlying problem is that the spike activity is naturally non-differentiable, raising more difficulties in supervised Training of SNNs. In this paper, we propose a spatio-temporal backpropagation (STBP) algorithm for Training high-performance spiking neural networks. In order to solve the non-differentiable problem of SNNs, an approximated derivative for spike activity is proposed, being appropriate for gradient descent Training. The STBP algorithm combines the layer-by-layer spatial domain (SD) and the timing-dependent temporal domain (TD), and does not require any additional complicated skill. We evaluate this method through adopting both the fully connected and convolutional architecture on the static MNIST dataset, a custom object detection dataset, and the dynamic N-MNIST dataset. Results bespeak that our approach achieves the best accuracy compared with existing state-of-the-art algorithms on spiking networks. This work provides a new perspective to investigate the high-performance SNNs for future brain-like computing paradigm with rich spatio-temporal dynamics.
Mark R Dixon - One of the best experts on this subject based on the ideXlab platform.
-
evaluating the relationships between the peak relational Training system Direct Training module assessment of basic language and learning skills revised and the vineland adaptive behavior scales ii
Journal of Developmental and Physical Disabilities, 2017Co-Authors: Albert Malkin, Mark R Dixon, Ryan C Speelman, Nicole LukeAbstract:The present study sought to examine the relationships between two language assessments and a psychometrically validated adaptive behavior scale. The assessments evaluated included the Promoting the Emergence of Advanced Knowledge Relational Training System - Direct Training Module (PEAK-DT), the Assessment of Basic Language and Learning Skills – Revised (ABLLS-R), and the Vineland Adaptive Behavior Scales, second edition (VABS-II). The assessments were completed for 21 children diagnosed with autism. Results indicate a significant correlation between scores on the PEAK-DT and ABLLS-R (r = 0.951, p < 0.001, PEAK-DT and VABS-II (r = 0.453, p < .05), as well as the ABLLS-R and VABS-II (r = 0.563, p < 0.05). The results did not indicate any ceiling effects amongst any of the assessments. These results extend research on the psychometric properties of these assessment tools and provide implications for practitioner choice of curricula.
-
normative sample of the peak relational Training system Direct Training module and subsequent comparisons to individuals with autism
Research in Autism Spectrum Disorders, 2014Co-Authors: Mark R Dixon, Jordan Belisle, Seth W Whiting, Kyle E RowseyAbstract:Abstract The present data provide a normative sample of the PEAK: Direct Training module assessment and a subsequent comparison to individuals with autism. Altogether, 206 typically developing participants and 94 participants with autism took part in the study. For the normative sample, there was a strong relationship between PEAK total score and age ( r = .659, p R 2 = .821, t = 18.51, p r = .021, p = .861), and that PEAK total scores for the autism group were significantly lower than the normative sample ( t (275) = 10.63, p
-
assessing the relationship between intelligence and the peak relational Training system
Research in Autism Spectrum Disorders, 2014Co-Authors: Mark R Dixon, Seth W Whiting, Kyle E Rowsey, Jordan BelislyAbstract:Abstract The Promoting the Emergence of Advanced Knowledge (PEAK) Relational Training System is an assessment and curriculum tool developed for basic and advanced skills using behavior analytic approaches. The current study evaluated the relationship between intelligence (as measured by IQ scores) and performance on the PEAK assessment with children with autism or other developmental and intellectual disabilities. Each child was administered the PEAK assessment from the Direct Training Module. Scores from this assessment were compared to IQ scores for all participants to assess the relationship between the two measures. Results indicated a strong, significant correlation between scores on standardized IQ tests and scores on the PEAK assessment (r = .759, p
-
peak relational Training system for children with autism and developmental disabilities correlations with peabody picture vocabulary test and assessment reliability
Journal of Developmental and Physical Disabilities, 2014Co-Authors: Mark R Dixon, Seth W Whiting, Josie Carman, Pamela A Tyler, Mary Rachel Enoch, Jacob H DaarAbstract:The present investigation sought to explore initial psychometric properties of the PEAK Relational Training System—Module 1: Direct Training for children with autism. Thirteen children diagnosed with autism or related disorders were exposed to an initial assessment designed to evaluate skill deficits within their repertoire, the Peabody Picture Vocabulary Test, and the Illinois Early Learning Standards Test. Additionally, staff performances were evaluated on reliability of delivery of the PEAK assessment. Results yielded significant positive correlations among the obtained PEAK assessment scores, the Peabody and the Standards assessments. Implications for evidence-based discrete trial Training curricula are discussed.
Luping Shi - One of the best experts on this subject based on the ideXlab platform.
-
Direct Training for spiking neural networks faster larger better
National Conference on Artificial Intelligence, 2019Co-Authors: Lei Deng, Jun Zhu, Yuan Xie, Luping ShiAbstract:Spiking neural networks (SNNs) that enables energy efficient implementation on emerging neuromorphic hardware are gaining more attention. Yet now, SNNs have not shown competitive performance compared with artificial neural networks (ANNs), due to the lack of effective learning algorithms and efficient programming frameworks. We address this issue from two aspects: (1) We propose a neuron normalization technique to adjust the neural selectivity and develop a Direct learning algorithm for deep SNNs. (2) Via narrowing the rate coding window and converting the leaky integrate-and-fire (LIF) model into an explicitly iterative version, we present a Pytorch-based implementation method towards the Training of large-scale SNNs. In this way, we are able to train deep SNNs with tens of times speedup. As a result, we achieve significantly better accuracy than the reported works on neuromorphic datasets (N-MNIST and DVSCIFAR10), and comparable accuracy as existing ANNs and pre-trained SNNs on non-spiking datasets (CIFAR10). To our best knowledge, this is the first work that demonstrates Direct Training of deep SNNs with high performance on CIFAR10, and the efficient implementation provides a new way to explore the potential of SNNs.
-
spatio temporal backpropagation for Training high performance spiking neural networks
Frontiers in Neuroscience, 2018Co-Authors: Lei Deng, Jun Zhu, Luping ShiAbstract:Spiking neural networks (SNNs) are promising in ascertaining brain-like behaviors since spikes are capable of encoding spatio-temporal information. Recent schemes, e.g. pre-Training from artificial neural networks (ANNs) or Direct Training based on backpropagation (BP), make the high-performance supervised Training of SNNs possible. However, these methods primarily fasten more attention on its spatial domain information, and the dynamics in temporal domain are attached less significance. Consequently, this might lead to the performance bottleneck, and scores of Training techniques shall be additionally required. Another underlying problem is that the spike activity is naturally non-differentiable, raising more difficulties in supervised Training of SNNs. In this paper, we propose a spatio-temporal backpropagation (STBP) algorithm for Training high-performance spiking neural networks. In order to solve the non-differentiable problem of SNNs, an approximated derivative for spike activity is proposed, being appropriate for gradient descent Training. The STBP algorithm combines the layer-by-layer spatial domain (SD) and the timing-dependent temporal domain (TD), and does not require any additional complicated skill. We evaluate this method through adopting both the fully connected and convolutional architecture on the static MNIST dataset, a custom object detection dataset, and the dynamic N-MNIST dataset. Results bespeak that our approach achieves the best accuracy compared with existing state-of-the-art algorithms on spiking networks. This work provides a new perspective to investigate the high-performance SNNs for future brain-like computing paradigm with rich spatio-temporal dynamics.
Kyle E Rowsey - One of the best experts on this subject based on the ideXlab platform.
-
normative sample of the peak relational Training system Direct Training module and subsequent comparisons to individuals with autism
Research in Autism Spectrum Disorders, 2014Co-Authors: Mark R Dixon, Jordan Belisle, Seth W Whiting, Kyle E RowseyAbstract:Abstract The present data provide a normative sample of the PEAK: Direct Training module assessment and a subsequent comparison to individuals with autism. Altogether, 206 typically developing participants and 94 participants with autism took part in the study. For the normative sample, there was a strong relationship between PEAK total score and age ( r = .659, p R 2 = .821, t = 18.51, p r = .021, p = .861), and that PEAK total scores for the autism group were significantly lower than the normative sample ( t (275) = 10.63, p
-
assessing the relationship between intelligence and the peak relational Training system
Research in Autism Spectrum Disorders, 2014Co-Authors: Mark R Dixon, Seth W Whiting, Kyle E Rowsey, Jordan BelislyAbstract:Abstract The Promoting the Emergence of Advanced Knowledge (PEAK) Relational Training System is an assessment and curriculum tool developed for basic and advanced skills using behavior analytic approaches. The current study evaluated the relationship between intelligence (as measured by IQ scores) and performance on the PEAK assessment with children with autism or other developmental and intellectual disabilities. Each child was administered the PEAK assessment from the Direct Training Module. Scores from this assessment were compared to IQ scores for all participants to assess the relationship between the two measures. Results indicated a strong, significant correlation between scores on standardized IQ tests and scores on the PEAK assessment (r = .759, p
Seth W Whiting - One of the best experts on this subject based on the ideXlab platform.
-
normative sample of the peak relational Training system Direct Training module and subsequent comparisons to individuals with autism
Research in Autism Spectrum Disorders, 2014Co-Authors: Mark R Dixon, Jordan Belisle, Seth W Whiting, Kyle E RowseyAbstract:Abstract The present data provide a normative sample of the PEAK: Direct Training module assessment and a subsequent comparison to individuals with autism. Altogether, 206 typically developing participants and 94 participants with autism took part in the study. For the normative sample, there was a strong relationship between PEAK total score and age ( r = .659, p R 2 = .821, t = 18.51, p r = .021, p = .861), and that PEAK total scores for the autism group were significantly lower than the normative sample ( t (275) = 10.63, p
-
assessing the relationship between intelligence and the peak relational Training system
Research in Autism Spectrum Disorders, 2014Co-Authors: Mark R Dixon, Seth W Whiting, Kyle E Rowsey, Jordan BelislyAbstract:Abstract The Promoting the Emergence of Advanced Knowledge (PEAK) Relational Training System is an assessment and curriculum tool developed for basic and advanced skills using behavior analytic approaches. The current study evaluated the relationship between intelligence (as measured by IQ scores) and performance on the PEAK assessment with children with autism or other developmental and intellectual disabilities. Each child was administered the PEAK assessment from the Direct Training Module. Scores from this assessment were compared to IQ scores for all participants to assess the relationship between the two measures. Results indicated a strong, significant correlation between scores on standardized IQ tests and scores on the PEAK assessment (r = .759, p
-
peak relational Training system for children with autism and developmental disabilities correlations with peabody picture vocabulary test and assessment reliability
Journal of Developmental and Physical Disabilities, 2014Co-Authors: Mark R Dixon, Seth W Whiting, Josie Carman, Pamela A Tyler, Mary Rachel Enoch, Jacob H DaarAbstract:The present investigation sought to explore initial psychometric properties of the PEAK Relational Training System—Module 1: Direct Training for children with autism. Thirteen children diagnosed with autism or related disorders were exposed to an initial assessment designed to evaluate skill deficits within their repertoire, the Peabody Picture Vocabulary Test, and the Illinois Early Learning Standards Test. Additionally, staff performances were evaluated on reliability of delivery of the PEAK assessment. Results yielded significant positive correlations among the obtained PEAK assessment scores, the Peabody and the Standards assessments. Implications for evidence-based discrete trial Training curricula are discussed.