Pub Date : 2022-08-17DOI: 10.1088/2634-4386/ac8a6a
Zirui Zhang, Dongliang Yang, Huihan Li, Ce Li, Zhongrui Wang, Linfeng Sun, Heejun Yang
Neuromorphic computing systems employing artificial synapses and neurons are expected to overcome the limitations of the present von Neumann computing architecture in terms of efficiency and bandwidth limits. Traditional neuromorphic devices have used 3D bulk materials, and thus, the resulting device size is difficult to be further scaled down for high density integration, which is required for highly integrated parallel computing. The emergence of two-dimensional (2D) materials offers a promising solution, as evidenced by the surge of reported 2D materials functioning as neuromorphic devices for next-generation computing. In this review, we summarize the 2D materials and their heterostructures to be used for neuromorphic computing devices, which could be classified by the working mechanism and device geometry. Then, we survey neuromorphic device arrays and their applications including artificial visual, tactile, and auditory functions. Finally, we discuss the current challenges of 2D materials to achieve practical neuromorphic devices, providing a perspective on the improved device performance, and integration level of the system. This will deepen our understanding of 2D materials and their heterojunctions and provide a guide to design highly performing memristors. At the same time, the challenges encountered in the industry are discussed, which provides a guide for the development direction of memristors.
{"title":"2D materials and van der Waals heterojunctions for neuromorphic computing","authors":"Zirui Zhang, Dongliang Yang, Huihan Li, Ce Li, Zhongrui Wang, Linfeng Sun, Heejun Yang","doi":"10.1088/2634-4386/ac8a6a","DOIUrl":"https://doi.org/10.1088/2634-4386/ac8a6a","url":null,"abstract":"Neuromorphic computing systems employing artificial synapses and neurons are expected to overcome the limitations of the present von Neumann computing architecture in terms of efficiency and bandwidth limits. Traditional neuromorphic devices have used 3D bulk materials, and thus, the resulting device size is difficult to be further scaled down for high density integration, which is required for highly integrated parallel computing. The emergence of two-dimensional (2D) materials offers a promising solution, as evidenced by the surge of reported 2D materials functioning as neuromorphic devices for next-generation computing. In this review, we summarize the 2D materials and their heterostructures to be used for neuromorphic computing devices, which could be classified by the working mechanism and device geometry. Then, we survey neuromorphic device arrays and their applications including artificial visual, tactile, and auditory functions. Finally, we discuss the current challenges of 2D materials to achieve practical neuromorphic devices, providing a perspective on the improved device performance, and integration level of the system. This will deepen our understanding of 2D materials and their heterojunctions and provide a guide to design highly performing memristors. At the same time, the challenges encountered in the industry are discussed, which provides a guide for the development direction of memristors.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"247 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123023489","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 : 2022-08-10DOI: 10.1088/2634-4386/ac889b
Xin Liu, Mingyu Yan, Lei Deng, Yujie Wu, De Han, Guoqi Li, Xiaochun Ye, Dongrui Fan
Emerging usages for millimeter wave (mmWave) radar have drawn extensive attention and inspired the exploration of learning mmWave radar data. To be effective, instead of using conventional approaches, recent works have employed modern neural network models to process mmWave radar data. However, due to some inevitable obstacles, e.g., noise and sparsity issues in data, the existing approaches are generally customized for specific scenarios. In this paper, we propose a general neuromorphic framework, termed mm-SNN, to process mmWave radar data with spiking neural networks (SNNs), leveraging the intrinsic advantages of SNNs in processing noisy and sparse data. Specifically, we first present the overall design of mm-SNN, which is adaptive and easily expanded for multi-sensor systems. Second, we introduce general and straightforward attention-based improvements into the mm-SNN to enhance the data representation, helping promote performance. Moreover, we conduct explorative experiments to certify the robustness and effectiveness of the mm-SNN. To the best of our knowledge, mm-SNN is the first SNN-based framework that processes mmWave radar data without using extra modules to alleviate the noise and sparsity issues, and at the same time, achieve considerable performance in the task of trajectory estimation.
{"title":"General spiking neural network framework for the learning trajectory from a noisy mmWave radar","authors":"Xin Liu, Mingyu Yan, Lei Deng, Yujie Wu, De Han, Guoqi Li, Xiaochun Ye, Dongrui Fan","doi":"10.1088/2634-4386/ac889b","DOIUrl":"https://doi.org/10.1088/2634-4386/ac889b","url":null,"abstract":"Emerging usages for millimeter wave (mmWave) radar have drawn extensive attention and inspired the exploration of learning mmWave radar data. To be effective, instead of using conventional approaches, recent works have employed modern neural network models to process mmWave radar data. However, due to some inevitable obstacles, e.g., noise and sparsity issues in data, the existing approaches are generally customized for specific scenarios. In this paper, we propose a general neuromorphic framework, termed mm-SNN, to process mmWave radar data with spiking neural networks (SNNs), leveraging the intrinsic advantages of SNNs in processing noisy and sparse data. Specifically, we first present the overall design of mm-SNN, which is adaptive and easily expanded for multi-sensor systems. Second, we introduce general and straightforward attention-based improvements into the mm-SNN to enhance the data representation, helping promote performance. Moreover, we conduct explorative experiments to certify the robustness and effectiveness of the mm-SNN. To the best of our knowledge, mm-SNN is the first SNN-based framework that processes mmWave radar data without using extra modules to alleviate the noise and sparsity issues, and at the same time, achieve considerable performance in the task of trajectory estimation.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134404047","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 : 2022-08-10DOI: 10.1088/2634-4386/ac889c
J. Aimone, Prasanna Date, Gabriel Andres Fonseca Guerra, Kathleen E. Hamilton, Kyle Henke, Bill Kay, G. Kenyon, Shruti R. Kulkarni, S. Mniszewski, Maryam Parsa, Sumedh R. Risbud, Catherine D. Schuman, William M. Severa, J. D. Smith
Though neuromorphic computers have typically targeted applications in machine learning and neuroscience (‘cognitive’ applications), they have many computational characteristics that are attractive for a wide variety of computational problems. In this work, we review the current state-of-the-art for non-cognitive applications on neuromorphic computers, including simple computational kernels for composition, graph algorithms, constrained optimization, and signal processing. We discuss the advantages of using neuromorphic computers for these different applications, as well as the challenges that still remain. The ultimate goal of this work is to bring awareness to this class of problems for neuromorphic systems to the broader community, particularly to encourage further work in this area and to make sure that these applications are considered in the design of future neuromorphic systems.
{"title":"A review of non-cognitive applications for neuromorphic computing","authors":"J. Aimone, Prasanna Date, Gabriel Andres Fonseca Guerra, Kathleen E. Hamilton, Kyle Henke, Bill Kay, G. Kenyon, Shruti R. Kulkarni, S. Mniszewski, Maryam Parsa, Sumedh R. Risbud, Catherine D. Schuman, William M. Severa, J. D. Smith","doi":"10.1088/2634-4386/ac889c","DOIUrl":"https://doi.org/10.1088/2634-4386/ac889c","url":null,"abstract":"Though neuromorphic computers have typically targeted applications in machine learning and neuroscience (‘cognitive’ applications), they have many computational characteristics that are attractive for a wide variety of computational problems. In this work, we review the current state-of-the-art for non-cognitive applications on neuromorphic computers, including simple computational kernels for composition, graph algorithms, constrained optimization, and signal processing. We discuss the advantages of using neuromorphic computers for these different applications, as well as the challenges that still remain. The ultimate goal of this work is to bring awareness to this class of problems for neuromorphic systems to the broader community, particularly to encourage further work in this area and to make sure that these applications are considered in the design of future neuromorphic systems.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129769465","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 : 2022-07-28DOI: 10.1088/2634-4386/ac84fd
Qingpeng Zhu, Chong Zhang, J. Triesch, Bertram E. Shi
The human eye has three rotational degrees of freedom: azimuthal, elevational, and torsional. Although torsional eye movements have the most limited excursion, Hering and Helmholtz have argued that they play an important role in optimizing visual information processing. In humans, the relationship between gaze direction and torsional eye angle is described by Listing’s law. However, it is still not clear how this behavior initially develops and remains calibrated during growth. Here we present the first computational model that enables an autonomous agent to learn and maintain binocular torsional eye movement control. In our model, two neural networks connected in series: one for sensory encoding followed by one for torsion control, are learned simultaneously as the agent behaves in the environment. Learning is based on the active efficient coding (AEC) framework, a generalization of Barlow’s efficient coding hypothesis to include action. Both networks adapt by minimizing the prediction error of the sensory representation, subject to a sparsity constraint on neural activity. The policies that emerge follow the predictions of Listing’s law. Because learning is driven by the sensorimotor contingencies experienced by the agent as it interacts with the environment, our system can adapt to the physical configuration of the agent as it changes. We propose that AEC provides the most parsimonious expression to date of Hering’s and Helmholtz’s hypotheses. We also demonstrate that it has practical implications in autonomous artificial vision systems, by providing an automatic and adaptive mechanism to correct orientation misalignments between cameras in a robotic active binocular vision head. Our system’s use of fairly low resolution (100 × 100 pixel) image windows and perceptual representations amenable to event-based input paves a pathway towards the implementation of adaptive self-calibrating robot control on neuromorphic hardware.
{"title":"Learning torsional eye movements through active efficient coding","authors":"Qingpeng Zhu, Chong Zhang, J. Triesch, Bertram E. Shi","doi":"10.1088/2634-4386/ac84fd","DOIUrl":"https://doi.org/10.1088/2634-4386/ac84fd","url":null,"abstract":"The human eye has three rotational degrees of freedom: azimuthal, elevational, and torsional. Although torsional eye movements have the most limited excursion, Hering and Helmholtz have argued that they play an important role in optimizing visual information processing. In humans, the relationship between gaze direction and torsional eye angle is described by Listing’s law. However, it is still not clear how this behavior initially develops and remains calibrated during growth. Here we present the first computational model that enables an autonomous agent to learn and maintain binocular torsional eye movement control. In our model, two neural networks connected in series: one for sensory encoding followed by one for torsion control, are learned simultaneously as the agent behaves in the environment. Learning is based on the active efficient coding (AEC) framework, a generalization of Barlow’s efficient coding hypothesis to include action. Both networks adapt by minimizing the prediction error of the sensory representation, subject to a sparsity constraint on neural activity. The policies that emerge follow the predictions of Listing’s law. Because learning is driven by the sensorimotor contingencies experienced by the agent as it interacts with the environment, our system can adapt to the physical configuration of the agent as it changes. We propose that AEC provides the most parsimonious expression to date of Hering’s and Helmholtz’s hypotheses. We also demonstrate that it has practical implications in autonomous artificial vision systems, by providing an automatic and adaptive mechanism to correct orientation misalignments between cameras in a robotic active binocular vision head. Our system’s use of fairly low resolution (100 × 100 pixel) image windows and perceptual representations amenable to event-based input paves a pathway towards the implementation of adaptive self-calibrating robot control on neuromorphic hardware.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"342 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123187391","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 : 2022-07-27DOI: 10.1088/2634-4386/acb2f0
Jinqi Huang, A. Serb, S. Stathopoulos, T. Prodromakis
Memristors, emerging non-volatile memory devices, have shown promising potential in neuromorphic hardware designs, especially in spiking neural network (SNN) hardware implementation. Memristor-based SNNs have been successfully applied in a wide range of applications, including image classification and pattern recognition. However, implementing memristor-based SNNs in text classification is still under exploration. One of the main reasons is that training memristor-based SNNs for text classification is costly due to the lack of efficient learning rules and memristor non-idealities. To address these issues and accelerate the research of exploring memristor-based SNNs in text classification applications, we develop a simulation framework with a virtual memristor array using an empirical memristor model. We use this framework to demonstrate a sentiment analysis task in the IMDB movie reviews dataset. We take two approaches to obtain trained SNNs with memristor models: (1) by converting a pre-trained artificial neural network (ANN) to a memristor-based SNN, or (2) by training a memristor-based SNN directly. These two approaches can be applied in two scenarios: offline classification and online training. We achieve the classification accuracy of 85.88% by converting a pre-trained ANN to a memristor-based SNN and 84.86% by training the memristor-based SNN directly, given that the baseline training accuracy of the equivalent ANN is 86.02%. We conclude that it is possible to achieve similar classification accuracy in simulation from ANNs to SNNs and from non-memristive synapses to data-driven memristive synapses. We also investigate how global parameters such as spike train length, the read noise, and the weight updating stop conditions affect the neural networks in both approaches. This investigation further indicates that the simulation using statistic memristor models in the two approaches presented by this paper can assist the exploration of memristor-based SNNs in natural language processing tasks.
{"title":"Text classification in memristor-based spiking neural networks","authors":"Jinqi Huang, A. Serb, S. Stathopoulos, T. Prodromakis","doi":"10.1088/2634-4386/acb2f0","DOIUrl":"https://doi.org/10.1088/2634-4386/acb2f0","url":null,"abstract":"Memristors, emerging non-volatile memory devices, have shown promising potential in neuromorphic hardware designs, especially in spiking neural network (SNN) hardware implementation. Memristor-based SNNs have been successfully applied in a wide range of applications, including image classification and pattern recognition. However, implementing memristor-based SNNs in text classification is still under exploration. One of the main reasons is that training memristor-based SNNs for text classification is costly due to the lack of efficient learning rules and memristor non-idealities. To address these issues and accelerate the research of exploring memristor-based SNNs in text classification applications, we develop a simulation framework with a virtual memristor array using an empirical memristor model. We use this framework to demonstrate a sentiment analysis task in the IMDB movie reviews dataset. We take two approaches to obtain trained SNNs with memristor models: (1) by converting a pre-trained artificial neural network (ANN) to a memristor-based SNN, or (2) by training a memristor-based SNN directly. These two approaches can be applied in two scenarios: offline classification and online training. We achieve the classification accuracy of 85.88% by converting a pre-trained ANN to a memristor-based SNN and 84.86% by training the memristor-based SNN directly, given that the baseline training accuracy of the equivalent ANN is 86.02%. We conclude that it is possible to achieve similar classification accuracy in simulation from ANNs to SNNs and from non-memristive synapses to data-driven memristive synapses. We also investigate how global parameters such as spike train length, the read noise, and the weight updating stop conditions affect the neural networks in both approaches. This investigation further indicates that the simulation using statistic memristor models in the two approaches presented by this paper can assist the exploration of memristor-based SNNs in natural language processing tasks.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"161 2","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114125057","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 : 2022-07-25DOI: 10.1088/2634-4386/ac94f3
Mohammadreza Mohammadi, Peyton S. Chandarana, J. Seekings, Sara Hendrix, Ramtin Zand
In this paper, we develop four spiking neural network (SNN) models for two static American sign language (ASL) hand gesture classification tasks, i.e., the ASL alphabet and ASL digits. The SNN models are deployed on Intel’s neuromorphic platform, Loihi, and then compared against equivalent deep neural network (DNN) models deployed on an edge computing device, the Intel neural compute stick 2 (NCS2). We perform a comprehensive comparison between the two systems in terms of accuracy, latency, power consumption, and energy. The best DNN model achieves an accuracy of 99.93% on the ASL alphabet dataset, whereas the best performing SNN model has an accuracy of 99.30%. For the ASL-digits dataset, the best DNN model achieves an accuracy of 99.76% accuracy while the SNN achieves 99.03%. Moreover, our obtained experimental results show that the Loihi neuromorphic hardware implementations achieve up to 20.64× and 4.10× reduction in power consumption and energy, respectively, when compared to NCS2.
{"title":"Static hand gesture recognition for American sign language using neuromorphic hardware","authors":"Mohammadreza Mohammadi, Peyton S. Chandarana, J. Seekings, Sara Hendrix, Ramtin Zand","doi":"10.1088/2634-4386/ac94f3","DOIUrl":"https://doi.org/10.1088/2634-4386/ac94f3","url":null,"abstract":"In this paper, we develop four spiking neural network (SNN) models for two static American sign language (ASL) hand gesture classification tasks, i.e., the ASL alphabet and ASL digits. The SNN models are deployed on Intel’s neuromorphic platform, Loihi, and then compared against equivalent deep neural network (DNN) models deployed on an edge computing device, the Intel neural compute stick 2 (NCS2). We perform a comprehensive comparison between the two systems in terms of accuracy, latency, power consumption, and energy. The best DNN model achieves an accuracy of 99.93% on the ASL alphabet dataset, whereas the best performing SNN model has an accuracy of 99.30%. For the ASL-digits dataset, the best DNN model achieves an accuracy of 99.76% accuracy while the SNN achieves 99.03%. Moreover, our obtained experimental results show that the Loihi neuromorphic hardware implementations achieve up to 20.64× and 4.10× reduction in power consumption and energy, respectively, when compared to NCS2.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122870853","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 : 2022-07-21DOI: 10.1088/2634-4386/ac830c
Mohammad Javad Mirshojaeian Hosseini, Yi Yang, Aidan J. Prendergast, Elisa Donati, M. Faezipour, G. Indiveri, Robert A. Nawrocki
In the nervous system synapses play a critical role in computation. In neuromorphic systems, biologically inspired hardware implementations of spiking neural networks, electronic synaptic circuits pass signals between silicon neurons by integrating pre-synaptic voltage pulses and converting them into post-synaptic currents, which are scaled by the synaptic weight parameter. The overwhelming majority of neuromorphic systems are implemented using inorganic, mainly silicon, technology. As such, they are physically rigid, require expensive fabrication equipment and high fabrication temperatures, are limited to small-area fabrication, and are difficult to interface with biological tissue. Organic electronics are based on electronic properties of carbon-based molecules and polymers and offer benefits including physical flexibility, low cost, low temperature, and large-area fabrication, as well as biocompatibility, all unavailable to inorganic electronics. Here, we demonstrate an organic differential-pair integrator synaptic circuit, a biologically realistic synapse model, implemented using physically flexible complementary organic electronics. The synapse is shown to convert input voltage spikes into output current traces with biologically realistic time scales. We characterize circuit’s responses based on various synaptic parameters, including gain and weighting voltages, time-constant, synaptic capacitance, and circuit response due to inputs of different frequencies. Time constants comparable to those of biological synapses and the neurons are critical in processing real-world sensory signals such as speech, or bio-signals measured from the body. For processing even slower signals, e.g., on behavioral time scales, we demonstrate time constants in excess of two seconds, while biologically plausible time constants are achieved by deploying smaller synaptic capacitors. We measure the circuit synaptic response to input voltage spikes and present the circuit response properties using custom-made circuit simulations, which are in good agreement with the measured behavior.
{"title":"An organic synaptic circuit: toward flexible and biocompatible organic neuromorphic processing","authors":"Mohammad Javad Mirshojaeian Hosseini, Yi Yang, Aidan J. Prendergast, Elisa Donati, M. Faezipour, G. Indiveri, Robert A. Nawrocki","doi":"10.1088/2634-4386/ac830c","DOIUrl":"https://doi.org/10.1088/2634-4386/ac830c","url":null,"abstract":"In the nervous system synapses play a critical role in computation. In neuromorphic systems, biologically inspired hardware implementations of spiking neural networks, electronic synaptic circuits pass signals between silicon neurons by integrating pre-synaptic voltage pulses and converting them into post-synaptic currents, which are scaled by the synaptic weight parameter. The overwhelming majority of neuromorphic systems are implemented using inorganic, mainly silicon, technology. As such, they are physically rigid, require expensive fabrication equipment and high fabrication temperatures, are limited to small-area fabrication, and are difficult to interface with biological tissue. Organic electronics are based on electronic properties of carbon-based molecules and polymers and offer benefits including physical flexibility, low cost, low temperature, and large-area fabrication, as well as biocompatibility, all unavailable to inorganic electronics. Here, we demonstrate an organic differential-pair integrator synaptic circuit, a biologically realistic synapse model, implemented using physically flexible complementary organic electronics. The synapse is shown to convert input voltage spikes into output current traces with biologically realistic time scales. We characterize circuit’s responses based on various synaptic parameters, including gain and weighting voltages, time-constant, synaptic capacitance, and circuit response due to inputs of different frequencies. Time constants comparable to those of biological synapses and the neurons are critical in processing real-world sensory signals such as speech, or bio-signals measured from the body. For processing even slower signals, e.g., on behavioral time scales, we demonstrate time constants in excess of two seconds, while biologically plausible time constants are achieved by deploying smaller synaptic capacitors. We measure the circuit synaptic response to input voltage spikes and present the circuit response properties using custom-made circuit simulations, which are in good agreement with the measured behavior.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121745722","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 : 2022-07-20DOI: 10.1088/2634-4386/acf1c5
Anmol Biswas, V. Saraswat, U. Ganguly
Spiking neural networks (SNNs) have emerged as a hardware efficient architecture for classification tasks. The challenge of spike-based encoding has been the lack of a universal training mechanism performed entirely using spikes. There have been several attempts to adopt the powerful backpropagation (BP) technique used in non-spiking artificial neural networks (ANNs): (1) SNNs can be trained by externally computed numerical gradients. (2) A major advancement towards native spike-based learning has been the use of approximate BP using spike-time dependent plasticity with phased forward/backward passes. However, the transfer of information between such phases for gradient and weight update calculation necessitates external memory and computational access. This is a challenge for standard neuromorphic hardware implementations. In this paper, we propose a stochastic SNN based back-prop (SSNN-BP) algorithm that utilizes a composite neuron to simultaneously compute the forward pass activations and backward pass gradients explicitly with spikes. Although signed gradient values are a challenge for spike-based representation, we tackle this by splitting the gradient signal into positive and negative streams. The composite neuron encodes information in the form of stochastic spike-trains and converts BP weight updates into temporally and spatially local spike coincidence updates compatible with hardware-friendly resistive processing units. Furthermore, we characterize the quantization effect of discrete spike-based weight update to show that our method approaches BP ANN baseline with sufficiently long spike-trains. Finally, we show that the well-performing softmax cross-entropy loss function can be implemented through inhibitory lateral connections enforcing a winner take all rule. Our SNN with a two-layer network shows excellent generalization through comparable performance to ANNs with equivalent architecture and regularization parameters on static image datasets like MNIST, Fashion-MNIST, Extended MNIST, and temporally encoded image datasets like Neuromorphic MNIST datasets. Thus, SSNN-BP enables BP compatible with purely spike-based neuromorphic hardware.
{"title":"A temporally and spatially local spike-based backpropagation algorithm to enable training in hardware","authors":"Anmol Biswas, V. Saraswat, U. Ganguly","doi":"10.1088/2634-4386/acf1c5","DOIUrl":"https://doi.org/10.1088/2634-4386/acf1c5","url":null,"abstract":"Spiking neural networks (SNNs) have emerged as a hardware efficient architecture for classification tasks. The challenge of spike-based encoding has been the lack of a universal training mechanism performed entirely using spikes. There have been several attempts to adopt the powerful backpropagation (BP) technique used in non-spiking artificial neural networks (ANNs): (1) SNNs can be trained by externally computed numerical gradients. (2) A major advancement towards native spike-based learning has been the use of approximate BP using spike-time dependent plasticity with phased forward/backward passes. However, the transfer of information between such phases for gradient and weight update calculation necessitates external memory and computational access. This is a challenge for standard neuromorphic hardware implementations. In this paper, we propose a stochastic SNN based back-prop (SSNN-BP) algorithm that utilizes a composite neuron to simultaneously compute the forward pass activations and backward pass gradients explicitly with spikes. Although signed gradient values are a challenge for spike-based representation, we tackle this by splitting the gradient signal into positive and negative streams. The composite neuron encodes information in the form of stochastic spike-trains and converts BP weight updates into temporally and spatially local spike coincidence updates compatible with hardware-friendly resistive processing units. Furthermore, we characterize the quantization effect of discrete spike-based weight update to show that our method approaches BP ANN baseline with sufficiently long spike-trains. Finally, we show that the well-performing softmax cross-entropy loss function can be implemented through inhibitory lateral connections enforcing a winner take all rule. Our SNN with a two-layer network shows excellent generalization through comparable performance to ANNs with equivalent architecture and regularization parameters on static image datasets like MNIST, Fashion-MNIST, Extended MNIST, and temporally encoded image datasets like Neuromorphic MNIST datasets. Thus, SSNN-BP enables BP compatible with purely spike-based neuromorphic hardware.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130041667","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 : 2022-07-20DOI: 10.1088/2634-4386/aca7de
C. Fields, K. Friston, J. Glazebrook, Michael Levin, A. Marcianò
We show how any finite physical system with morphological, i.e. three-dimensional embedding or shape, degrees of freedom and locally limited free energy will, under the constraints of the free energy principle, evolve over time towards a neuromorphic morphology that supports hierarchical computations in which each ‘level’ of the hierarchy enacts a coarse-graining of its inputs, and dually, a fine-graining of its outputs. Such hierarchies occur throughout biology, from the architectures of intracellular signal transduction pathways to the large-scale organization of perception and action cycles in the mammalian brain. The close formal connections between cone-cocone diagrams (CCCD) as models of quantum reference frames on the one hand, and between CCCDs and topological quantum field theories on the other, allow the representation of such computations in the fully-general quantum-computational framework of topological quantum neural networks.
{"title":"The free energy principle induces neuromorphic development","authors":"C. Fields, K. Friston, J. Glazebrook, Michael Levin, A. Marcianò","doi":"10.1088/2634-4386/aca7de","DOIUrl":"https://doi.org/10.1088/2634-4386/aca7de","url":null,"abstract":"We show how any finite physical system with morphological, i.e. three-dimensional embedding or shape, degrees of freedom and locally limited free energy will, under the constraints of the free energy principle, evolve over time towards a neuromorphic morphology that supports hierarchical computations in which each ‘level’ of the hierarchy enacts a coarse-graining of its inputs, and dually, a fine-graining of its outputs. Such hierarchies occur throughout biology, from the architectures of intracellular signal transduction pathways to the large-scale organization of perception and action cycles in the mammalian brain. The close formal connections between cone-cocone diagrams (CCCD) as models of quantum reference frames on the one hand, and between CCCDs and topological quantum field theories on the other, allow the representation of such computations in the fully-general quantum-computational framework of topological quantum neural networks.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"126 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128078350","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 : 2022-07-19DOI: 10.1088/2634-4386/ac9e6f
Zhaoqi Chen, Alia Nasrallah, Milad Alemohammad, Masanori Furuta, R. Etienne-Cummings
In this paper, we propose a simplified and robust model for place cell generation based on the oscillatory interference (OI) model concept. Aiming toward hardware implementation in bio-inspired simultaneous localization and mapping (SLAM) systems for mobile robotics, we base our model on logic operations that reduce its computational complexity. The model compensates for parameter variations in the behaviors of the population of constituent theta cells, and allows the theta cells to have square-wave oscillation profiles. The robustness of the model, with respect to mismatch in the theta cell’s base oscillation frequency and gain—as a function of modulatory inputs—is demonstrated. Place cell composed of 48 theta cells with base frequency variations with a 25% standard deviation from the mean and a gain error with 20% standard deviation from the mean only result in a 20% deformations within the place field and 0.24% outer side lobes, and an overall pattern with 0.0015 mean squared error on average. We also present how the model can be used to achieve the localization and path-tracking functionalities of SLAM. Hence, we propose a model for spatial cell formation using theta cells with behaviors that are biologically plausible and hardware implementable for real world application in neurally-inspired SLAM.
{"title":"Neuromorphic model of hippocampus place cells using an oscillatory interference technique for hardware implementation","authors":"Zhaoqi Chen, Alia Nasrallah, Milad Alemohammad, Masanori Furuta, R. Etienne-Cummings","doi":"10.1088/2634-4386/ac9e6f","DOIUrl":"https://doi.org/10.1088/2634-4386/ac9e6f","url":null,"abstract":"In this paper, we propose a simplified and robust model for place cell generation based on the oscillatory interference (OI) model concept. Aiming toward hardware implementation in bio-inspired simultaneous localization and mapping (SLAM) systems for mobile robotics, we base our model on logic operations that reduce its computational complexity. The model compensates for parameter variations in the behaviors of the population of constituent theta cells, and allows the theta cells to have square-wave oscillation profiles. The robustness of the model, with respect to mismatch in the theta cell’s base oscillation frequency and gain—as a function of modulatory inputs—is demonstrated. Place cell composed of 48 theta cells with base frequency variations with a 25% standard deviation from the mean and a gain error with 20% standard deviation from the mean only result in a 20% deformations within the place field and 0.24% outer side lobes, and an overall pattern with 0.0015 mean squared error on average. We also present how the model can be used to achieve the localization and path-tracking functionalities of SLAM. Hence, we propose a model for spatial cell formation using theta cells with behaviors that are biologically plausible and hardware implementable for real world application in neurally-inspired SLAM.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130258767","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：