Pub Date : 2023-03-16DOI: 10.1088/2634-4386/ace737
Jonathan Timcheck, S. Shrestha, D. B. Rubin, A. Kupryjanow, G. Orchard, Lukasz Pindor, Timothy M. Shea, Mike Davies
A critical enabler for progress in neuromorphic computing research is the ability to transparently evaluate different neuromorphic solutions on important tasks and to compare them to state-of-the-art conventional solutions. The Intel Neuromorphic Deep Noise Suppression Challenge (Intel N-DNS Challenge), inspired by the Microsoft DNS Challenge, tackles a ubiquitous and commercially relevant task: real-time audio denoising. Audio denoising is likely to reap the benefits of neuromorphic computing due to its low-bandwidth, temporal nature and its relevance for low-power devices. The Intel N-DNS Challenge consists of two tracks: a simulation-based algorithmic track to encourage algorithmic innovation, and a neuromorphic hardware (Loihi 2) track to rigorously evaluate solutions. For both tracks, we specify an evaluation methodology based on energy, latency, and resource consumption in addition to output audio quality. We make the Intel N-DNS Challenge dataset scripts and evaluation code freely accessible, encourage community participation with monetary prizes, and release a neuromorphic baseline solution which shows promising audio quality, high power efficiency, and low resource consumption when compared to Microsoft NsNet2 and a proprietary Intel denoising model used in production. We hope the Intel N-DNS Challenge will hasten innovation in neuromorphic algorithms research, especially in the area of training tools and methods for real-time signal processing. We expect the winners of the challenge will demonstrate that for problems like audio denoising, significant gains in power and resources can be realized on neuromorphic devices available today compared to conventional state-of-the-art solutions.
{"title":"The Intel neuromorphic DNS challenge","authors":"Jonathan Timcheck, S. Shrestha, D. B. Rubin, A. Kupryjanow, G. Orchard, Lukasz Pindor, Timothy M. Shea, Mike Davies","doi":"10.1088/2634-4386/ace737","DOIUrl":"https://doi.org/10.1088/2634-4386/ace737","url":null,"abstract":"A critical enabler for progress in neuromorphic computing research is the ability to transparently evaluate different neuromorphic solutions on important tasks and to compare them to state-of-the-art conventional solutions. The Intel Neuromorphic Deep Noise Suppression Challenge (Intel N-DNS Challenge), inspired by the Microsoft DNS Challenge, tackles a ubiquitous and commercially relevant task: real-time audio denoising. Audio denoising is likely to reap the benefits of neuromorphic computing due to its low-bandwidth, temporal nature and its relevance for low-power devices. The Intel N-DNS Challenge consists of two tracks: a simulation-based algorithmic track to encourage algorithmic innovation, and a neuromorphic hardware (Loihi 2) track to rigorously evaluate solutions. For both tracks, we specify an evaluation methodology based on energy, latency, and resource consumption in addition to output audio quality. We make the Intel N-DNS Challenge dataset scripts and evaluation code freely accessible, encourage community participation with monetary prizes, and release a neuromorphic baseline solution which shows promising audio quality, high power efficiency, and low resource consumption when compared to Microsoft NsNet2 and a proprietary Intel denoising model used in production. We hope the Intel N-DNS Challenge will hasten innovation in neuromorphic algorithms research, especially in the area of training tools and methods for real-time signal processing. We expect the winners of the challenge will demonstrate that for problems like audio denoising, significant gains in power and resources can be realized on neuromorphic devices available today compared to conventional state-of-the-art solutions.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"23 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130229979","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 : 2023-03-09DOI: 10.1088/2634-4386/acc2e1
Silvia Tolu, Beck Strohmer, Omar Zahra
Neurorobotics has emerged from the alliance between neuroscience and robotics. It pursues the investigation of reproducing living organism-like behaviors in robots by means of the embodiment of computational models of the central nervous system. This perspective article discusses the current trend of implementing tools for the pressing challenge of early-diagnosis of neurodegenerative diseases and how neurorobotics approaches can help. Recently, advances in this field have allowed the testing of some neuroscientific hypotheses related to brain diseases, but the lack of biological plausibility of developed brain models and musculoskeletal systems has limited the understanding of the underlying brain mechanisms that lead to deficits in motor and cognitive tasks. Key aspects and methods to enhance the reproducibility of natural behaviors observed in healthy and impaired brains are proposed in this perspective. In the long term, the goal is to move beyond finding therapies and look into how researchers can use neurorobotics to reduce testing on humans as well as find root causes for disease.
{"title":"Perspective on investigation of neurodegenerative diseases with neurorobotics approaches","authors":"Silvia Tolu, Beck Strohmer, Omar Zahra","doi":"10.1088/2634-4386/acc2e1","DOIUrl":"https://doi.org/10.1088/2634-4386/acc2e1","url":null,"abstract":"Neurorobotics has emerged from the alliance between neuroscience and robotics. It pursues the investigation of reproducing living organism-like behaviors in robots by means of the embodiment of computational models of the central nervous system. This perspective article discusses the current trend of implementing tools for the pressing challenge of early-diagnosis of neurodegenerative diseases and how neurorobotics approaches can help. Recently, advances in this field have allowed the testing of some neuroscientific hypotheses related to brain diseases, but the lack of biological plausibility of developed brain models and musculoskeletal systems has limited the understanding of the underlying brain mechanisms that lead to deficits in motor and cognitive tasks. Key aspects and methods to enhance the reproducibility of natural behaviors observed in healthy and impaired brains are proposed in this perspective. In the long term, the goal is to move beyond finding therapies and look into how researchers can use neurorobotics to reduce testing on humans as well as find root causes for disease.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133399283","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 : 2023-03-03DOI: 10.1088/2634-4386/acdb96
Matthew O. A. Ellis, A. Welbourne, Stephan J. Kyle, P. Fry, D. Allwood, T. Hayward, E. Vasilaki
The impressive performance of artificial neural networks has come at the cost of high energy usage and CO2 emissions. Unconventional computing architectures, with magnetic systems as a candidate, have potential as alternative energy-efficient hardware, but, still face challenges, such as stochastic behaviour, in implementation. Here, we present a methodology for exploiting the traditionally detrimental stochastic effects in magnetic domain-wall motion in nanowires. We demonstrate functional binary stochastic synapses alongside a gradient learning rule that allows their training with applicability to a range of stochastic systems. The rule, utilising the mean and variance of the neuronal output distribution, finds a trade-off between synaptic stochasticity and energy efficiency depending on the number of measurements of each synapse. For single measurements, the rule results in binary synapses with minimal stochasticity, sacrificing potential performance for robustness. For multiple measurements, synaptic distributions are broad, approximating better-performing continuous synapses. This observation allows us to choose design principles depending on the desired performance and the device’s operational speed and energy cost. We verify performance on physical hardware, showing it is comparable to a standard neural network.
{"title":"Machine learning using magnetic stochastic synapses","authors":"Matthew O. A. Ellis, A. Welbourne, Stephan J. Kyle, P. Fry, D. Allwood, T. Hayward, E. Vasilaki","doi":"10.1088/2634-4386/acdb96","DOIUrl":"https://doi.org/10.1088/2634-4386/acdb96","url":null,"abstract":"The impressive performance of artificial neural networks has come at the cost of high energy usage and CO2 emissions. Unconventional computing architectures, with magnetic systems as a candidate, have potential as alternative energy-efficient hardware, but, still face challenges, such as stochastic behaviour, in implementation. Here, we present a methodology for exploiting the traditionally detrimental stochastic effects in magnetic domain-wall motion in nanowires. We demonstrate functional binary stochastic synapses alongside a gradient learning rule that allows their training with applicability to a range of stochastic systems. The rule, utilising the mean and variance of the neuronal output distribution, finds a trade-off between synaptic stochasticity and energy efficiency depending on the number of measurements of each synapse. For single measurements, the rule results in binary synapses with minimal stochasticity, sacrificing potential performance for robustness. For multiple measurements, synaptic distributions are broad, approximating better-performing continuous synapses. This observation allows us to choose design principles depending on the desired performance and the device’s operational speed and energy cost. We verify performance on physical hardware, showing it is comparable to a standard neural network.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130981016","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 : 2023-03-02DOI: 10.1088/2634-4386/acc08e
Zahra Yousefi Darani, Iacopo Hachen, M. E. Diamond
In the future, artificial agents will need to make assessments of tactile stimuli in order to interact intelligently with the environment and with humans. Such assessments will depend on exquisite and robust mechanosensors, but sensors alone do not make judgments and choices. Rather, the central processing of mechanosensor inputs must be implemented with algorithms that produce ‘behavioral states’ in the artificial agent that resemble or mimic perceptual judgments in biology. In this study, we consider the problem of perceptual judgment as applied to vibration intensity. By a combination of computational modeling and simulation followed by psychophysical testing of vibration intensity perception in rats, we show that a simple yet highly salient judgment—is the current stimulus strong or weak?—can be explained as the comparison of ongoing sensory input against a criterion constructed as the time-weighted average of the history of recent stimuli. Simulations and experiments explore how judgments are shaped by the distribution of stimuli along the intensity dimension and, most importantly, by the time constant of integration which dictates the dynamics of criterion updating. The findings of this study imply that judgments made by the real nervous system are not absolute readouts of physical parameters but are context-dependent; algorithms of this form can be built into artificial systems.
{"title":"Dynamics of the judgment of tactile stimulus intensity","authors":"Zahra Yousefi Darani, Iacopo Hachen, M. E. Diamond","doi":"10.1088/2634-4386/acc08e","DOIUrl":"https://doi.org/10.1088/2634-4386/acc08e","url":null,"abstract":"In the future, artificial agents will need to make assessments of tactile stimuli in order to interact intelligently with the environment and with humans. Such assessments will depend on exquisite and robust mechanosensors, but sensors alone do not make judgments and choices. Rather, the central processing of mechanosensor inputs must be implemented with algorithms that produce ‘behavioral states’ in the artificial agent that resemble or mimic perceptual judgments in biology. In this study, we consider the problem of perceptual judgment as applied to vibration intensity. By a combination of computational modeling and simulation followed by psychophysical testing of vibration intensity perception in rats, we show that a simple yet highly salient judgment—is the current stimulus strong or weak?—can be explained as the comparison of ongoing sensory input against a criterion constructed as the time-weighted average of the history of recent stimuli. Simulations and experiments explore how judgments are shaped by the distribution of stimuli along the intensity dimension and, most importantly, by the time constant of integration which dictates the dynamics of criterion updating. The findings of this study imply that judgments made by the real nervous system are not absolute readouts of physical parameters but are context-dependent; algorithms of this form can be built into artificial systems.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"15 2","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120847565","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 : 2023-03-01DOI: 10.1088/2634-4386/acc050
Juan Wen, Zhen-Ye Zhu, Xin Guo
The human visual system encodes optical information perceived by photoreceptors in the retina into neural spikes and then processes them by the visual cortex, with high efficiency and low energy consumption. Inspired by this information processing mode, an universal artificial neuron constructed with a resistor (R s) and a threshold switching memristor can realize rate coding by modulating pulse parameters and the resistance of R s. Owing to the absence of an external parallel capacitor, the artificial neuron has minimized chip area. In addition, an artificial visual neuron is proposed by replacing R s in the artificial neuron with a photo-resistor. The oscillation frequency of the artificial visual neuron depends on the distance between the photo-resistor and light, which is fundamental to acquiring depth perception for precise recognition and learning. A visual perception system with the artificial visual neuron can accurately and conceptually emulate the self-regulation process of the speed control system in a driverless automobile. Therefore, the artificial visual neuron can process efficiently sensory data, reduce or eliminate data transfer and conversion at sensor/processor interfaces, and expand its application in the field of artificial intelligence.
{"title":"Artificial visual neuron based on threshold switching memristors","authors":"Juan Wen, Zhen-Ye Zhu, Xin Guo","doi":"10.1088/2634-4386/acc050","DOIUrl":"https://doi.org/10.1088/2634-4386/acc050","url":null,"abstract":"The human visual system encodes optical information perceived by photoreceptors in the retina into neural spikes and then processes them by the visual cortex, with high efficiency and low energy consumption. Inspired by this information processing mode, an universal artificial neuron constructed with a resistor (R s) and a threshold switching memristor can realize rate coding by modulating pulse parameters and the resistance of R s. Owing to the absence of an external parallel capacitor, the artificial neuron has minimized chip area. In addition, an artificial visual neuron is proposed by replacing R s in the artificial neuron with a photo-resistor. The oscillation frequency of the artificial visual neuron depends on the distance between the photo-resistor and light, which is fundamental to acquiring depth perception for precise recognition and learning. A visual perception system with the artificial visual neuron can accurately and conceptually emulate the self-regulation process of the speed control system in a driverless automobile. Therefore, the artificial visual neuron can process efficiently sensory data, reduce or eliminate data transfer and conversion at sensor/processor interfaces, and expand its application in the field of artificial intelligence.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131488684","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 : 2023-03-01DOI: 10.1088/2634-4386/acc04f
N. Szczecinski, C. Goldsmith, W. Nourse, R. Quinn
This article is a historical perspective on how the study of the neuromechanics of insects and other arthropods has inspired the construction, and especially the control, of hexapod robots. Many hexapod robots’ control systems share common features, including: 1. Direction of motor output of each joint (i.e. to flex or extend) in the leg is gated by an oscillatory or bistable gating mechanism; 2. The relative phasing between each joint is influenced by proprioceptive feedback from the periphery (e.g. joint angles, leg load) or central connections between joint controllers; and 3. Behavior can be directed (e.g. transition from walking along a straight path to walking along a curve) via low-dimensional, broadly-acting descending inputs to the network. These distributed control schemes are inspired by, and in some robots, closely mimic the organization of the nervous systems of insects, the natural hexapods, as well as crustaceans. Nearly a century of research has revealed organizational principles such as central pattern generators, the role of proprioceptive feedback in control, and command neurons. These concepts have inspired the control systems of hexapod robots in the past, in which these structures were applied to robot controllers with neuromorphic (i.e. distributed) organization, but not neuromorphic computational units (i.e. neurons) or computational hardware (i.e. hardware-accelerated neurons). Presently, several hexapod robots are controlled with neuromorphic computational units with or without neuromorphic organization, almost always without neuromorphic hardware. In the near future, we expect to see hexapod robots whose controllers include neuromorphic organization, computational units, and hardware. Such robots may exhibit the full mobility of their insect counterparts thanks to a ‘biology-first’ approach to controller design. This perspective article is not a comprehensive review of the neuroscientific literature but is meant to give those with engineering backgrounds a gentle introduction into the neuroscientific principles that underlie models and inspire neuromorphic robot controllers. A historical summary of hexapod robots whose control systems and behaviors use neuromorphic elements is provided. Robots whose controllers closely model animals and may be used to generate concrete hypotheses for future animal experiments are of particular interest to the authors. The authors hope that by highlighting the decades of experimental research that has led to today’s accepted organization principles of arthropod nervous systems, engineers may better understand these systems and more fully apply biological details in their robots. To assist the interested reader, deeper reviews of particular topics from biology are suggested throughout.
{"title":"A perspective on the neuromorphic control of legged locomotion in past, present, and future insect-like robots","authors":"N. Szczecinski, C. Goldsmith, W. Nourse, R. Quinn","doi":"10.1088/2634-4386/acc04f","DOIUrl":"https://doi.org/10.1088/2634-4386/acc04f","url":null,"abstract":"This article is a historical perspective on how the study of the neuromechanics of insects and other arthropods has inspired the construction, and especially the control, of hexapod robots. Many hexapod robots’ control systems share common features, including: 1. Direction of motor output of each joint (i.e. to flex or extend) in the leg is gated by an oscillatory or bistable gating mechanism; 2. The relative phasing between each joint is influenced by proprioceptive feedback from the periphery (e.g. joint angles, leg load) or central connections between joint controllers; and 3. Behavior can be directed (e.g. transition from walking along a straight path to walking along a curve) via low-dimensional, broadly-acting descending inputs to the network. These distributed control schemes are inspired by, and in some robots, closely mimic the organization of the nervous systems of insects, the natural hexapods, as well as crustaceans. Nearly a century of research has revealed organizational principles such as central pattern generators, the role of proprioceptive feedback in control, and command neurons. These concepts have inspired the control systems of hexapod robots in the past, in which these structures were applied to robot controllers with neuromorphic (i.e. distributed) organization, but not neuromorphic computational units (i.e. neurons) or computational hardware (i.e. hardware-accelerated neurons). Presently, several hexapod robots are controlled with neuromorphic computational units with or without neuromorphic organization, almost always without neuromorphic hardware. In the near future, we expect to see hexapod robots whose controllers include neuromorphic organization, computational units, and hardware. Such robots may exhibit the full mobility of their insect counterparts thanks to a ‘biology-first’ approach to controller design. This perspective article is not a comprehensive review of the neuroscientific literature but is meant to give those with engineering backgrounds a gentle introduction into the neuroscientific principles that underlie models and inspire neuromorphic robot controllers. A historical summary of hexapod robots whose control systems and behaviors use neuromorphic elements is provided. Robots whose controllers closely model animals and may be used to generate concrete hypotheses for future animal experiments are of particular interest to the authors. The authors hope that by highlighting the decades of experimental research that has led to today’s accepted organization principles of arthropod nervous systems, engineers may better understand these systems and more fully apply biological details in their robots. To assist the interested reader, deeper reviews of particular topics from biology are suggested throughout.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"9 5","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120845665","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 : 2023-02-28DOI: 10.1088/2634-4386/acbfd6
S. Prosandeev, S. Prokhorenko, Y. Nahas, Yali Yang, Changsong Xu, J. Grollier, D. Talbayev, B. Dkhil, L. Bellaiche
This review summarizes recent works, all using a specific atomistic approach, that predict and explain the occurrence of key features for neuromorphic computing in three archetypical dipolar materials, when they are subject to THz excitations. The main ideas behind such atomistic approach are provided, and illustration of model relaxor ferroelectrics, antiferroelectrics, and normal ferroelectrics are given, highlighting the important potential of polar materials as candidates for neuromorphic computing. Some peculiar emphases are made in this Review, such as the connection between neuromorphic features and percolation theory, local minima in energy path, topological transitions and/or anharmonic oscillator model, depending on the material under investigation. By considering three different and main polar material families, this work provides a complete and innovative toolbox for designing polar-based neuromorphic systems.
{"title":"Designing polar textures with ultrafast neuromorphic features from atomistic simulations","authors":"S. Prosandeev, S. Prokhorenko, Y. Nahas, Yali Yang, Changsong Xu, J. Grollier, D. Talbayev, B. Dkhil, L. Bellaiche","doi":"10.1088/2634-4386/acbfd6","DOIUrl":"https://doi.org/10.1088/2634-4386/acbfd6","url":null,"abstract":"This review summarizes recent works, all using a specific atomistic approach, that predict and explain the occurrence of key features for neuromorphic computing in three archetypical dipolar materials, when they are subject to THz excitations. The main ideas behind such atomistic approach are provided, and illustration of model relaxor ferroelectrics, antiferroelectrics, and normal ferroelectrics are given, highlighting the important potential of polar materials as candidates for neuromorphic computing. Some peculiar emphases are made in this Review, such as the connection between neuromorphic features and percolation theory, local minima in energy path, topological transitions and/or anharmonic oscillator model, depending on the material under investigation. By considering three different and main polar material families, this work provides a complete and innovative toolbox for designing polar-based neuromorphic systems.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115440540","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 : 2023-02-09DOI: 10.1088/2634-4386/acbab8
Yikai Yang, J. K. Eshraghian, Nhan Duy Truong, A. Nikpour, O. Kavehei
The vast majority of studies that process and analyze neural signals are conducted on cloud computing resources, which is often necessary for the demanding requirements of deep neural network workloads. However, applications such as epileptic seizure detection stand to benefit from edge devices that can securely analyze sensitive medical data in a real-time and personalised manner. In this work, we propose a novel neuromorphic computing approach to seizure detection using a surrogate gradient-based deep spiking neural network (SNN), which consists of a novel spiking ConvLSTM unit. We have trained, validated, and rigorously tested the proposed SNN model across three publicly accessible datasets, including Boston Children’s Hospital–MIT (CHB-MIT) dataset from the U.S., and the Freiburg (FB) and EPILEPSIAE intracranial electroencephalogram datasets from Germany. The average leave-one-out cross-validation area under the curve score for FB, CHB-MIT and EPILEPSIAE datasets can reach 92.7 % , 89.0 % , and 81.1 % , respectively, while the computational overhead and energy consumption are significantly reduced when compared to alternative state-of-the-art models, showing the potential for building an accurate hardware-friendly, low-power neuromorphic system. This is the first feasibility study using a deep SNN for seizure detection on several reliable public datasets.
{"title":"Neuromorphic deep spiking neural networks for seizure detection","authors":"Yikai Yang, J. K. Eshraghian, Nhan Duy Truong, A. Nikpour, O. Kavehei","doi":"10.1088/2634-4386/acbab8","DOIUrl":"https://doi.org/10.1088/2634-4386/acbab8","url":null,"abstract":"The vast majority of studies that process and analyze neural signals are conducted on cloud computing resources, which is often necessary for the demanding requirements of deep neural network workloads. However, applications such as epileptic seizure detection stand to benefit from edge devices that can securely analyze sensitive medical data in a real-time and personalised manner. In this work, we propose a novel neuromorphic computing approach to seizure detection using a surrogate gradient-based deep spiking neural network (SNN), which consists of a novel spiking ConvLSTM unit. We have trained, validated, and rigorously tested the proposed SNN model across three publicly accessible datasets, including Boston Children’s Hospital–MIT (CHB-MIT) dataset from the U.S., and the Freiburg (FB) and EPILEPSIAE intracranial electroencephalogram datasets from Germany. The average leave-one-out cross-validation area under the curve score for FB, CHB-MIT and EPILEPSIAE datasets can reach 92.7 % , 89.0 % , and 81.1 % , respectively, while the computational overhead and energy consumption are significantly reduced when compared to alternative state-of-the-art models, showing the potential for building an accurate hardware-friendly, low-power neuromorphic system. This is the first feasibility study using a deep SNN for seizure detection on several reliable public datasets.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122151792","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}
Deep neural networks (DNNs) are one of the key fields of machine learning. It requires considerable computational resources for cognitive tasks. As a novel technology to perform computing inside/near memory units, in-memory computing (IMC) significantly improves computing efficiency by reducing the need for repetitive data transfer between the processing and memory units. However, prior IMC designs mainly focus on the acceleration for DNN inference. DNN training with the IMC hardware has rarely been proposed. The challenges lie in the requirement of DNN training for high precision (e.g. floating point (FP)) and various operations of tensors (e.g. inner and outer products). These challenges call for the IMC design with new features. This paper proposes a novel Hadamard product-based IMC design for FP DNN training. Our design consists of multiple compartments, which are the basic units for the matrix element-wise processing. We also develop BFloat16 post-processing circuits and fused adder trees, laying the foundation for IMC FP processing. Based on the proposed circuit scheme, we reformulate the back-propagation training algorithm for the convenience and efficiency of the IMC execution. The proposed design is implemented with commercial 28 nm technology process design kits and benchmarked with widely used neural networks. We model the influence of the circuit structural design parameters and provide an analysis framework for design space exploration. Our simulation validates that MobileNet training with the proposed IMC scheme saves 91.2% in energy and 13.9% in time versus the same task with NVIDIA GTX 3060 GPU. The proposed IMC design has a data density of 769.2 Kb mm−2 with the FP processing circuits included, showing a 3.5 × improvement than the prior FP IMC designs.
{"title":"Hadamard product-based in-memory computing design for floating point neural network training","authors":"Anjunyi Fan, Yihan Fu, Yaoyu Tao, Zhonghua Jin, Haiyue Han, Huiyu Liu, Yaojun Zhang, Bonan Yan, Yuch-Chi Yang, Ru Huang","doi":"10.1088/2634-4386/acbab9","DOIUrl":"https://doi.org/10.1088/2634-4386/acbab9","url":null,"abstract":"Deep neural networks (DNNs) are one of the key fields of machine learning. It requires considerable computational resources for cognitive tasks. As a novel technology to perform computing inside/near memory units, in-memory computing (IMC) significantly improves computing efficiency by reducing the need for repetitive data transfer between the processing and memory units. However, prior IMC designs mainly focus on the acceleration for DNN inference. DNN training with the IMC hardware has rarely been proposed. The challenges lie in the requirement of DNN training for high precision (e.g. floating point (FP)) and various operations of tensors (e.g. inner and outer products). These challenges call for the IMC design with new features. This paper proposes a novel Hadamard product-based IMC design for FP DNN training. Our design consists of multiple compartments, which are the basic units for the matrix element-wise processing. We also develop BFloat16 post-processing circuits and fused adder trees, laying the foundation for IMC FP processing. Based on the proposed circuit scheme, we reformulate the back-propagation training algorithm for the convenience and efficiency of the IMC execution. The proposed design is implemented with commercial 28 nm technology process design kits and benchmarked with widely used neural networks. We model the influence of the circuit structural design parameters and provide an analysis framework for design space exploration. Our simulation validates that MobileNet training with the proposed IMC scheme saves 91.2% in energy and 13.9% in time versus the same task with NVIDIA GTX 3060 GPU. The proposed IMC design has a data density of 769.2 Kb mm−2 with the FP processing circuits included, showing a 3.5 × improvement than the prior FP IMC designs.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"98 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127713107","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 : 2023-02-08DOI: 10.1088/2634-4386/acba3f
Feng Miao, J. JoshuaYang, I. Valov, Yang Chai
Neuromorphic computing aims at mimicking the synapses, dendrites, and neurons in the brain as well as their associated connected networks to perform a variety of complex tasks including sensing, computing, perception, sometimes by directly utilizing the physical properties of materials. Their functionality diversity and performance highly depend on the use of materials. Compared to the conventional materials, 2D materials exhibit many unique physical properties and the research of 2D materials has reshaped the field of neuromorphic computing. This special issue presents some of the innovations in using devices based on 2D materials to emulate the biological synapses or generate noise injection to hardware neural networks. The issue also provides a comprehensive analysis of recent advances in exploiting the unique physical properties of 2D materials for neuromorphic computing. These innovations and analysis may serve as a useful guide to further advance 2D materials for practical applications. This special issue includes two research articles and four review articles, with contents briefly summarized in the following paragraphs.
{"title":"Editorial: Focus issue on 2D materials for neuromorphic computing","authors":"Feng Miao, J. JoshuaYang, I. Valov, Yang Chai","doi":"10.1088/2634-4386/acba3f","DOIUrl":"https://doi.org/10.1088/2634-4386/acba3f","url":null,"abstract":"\u0000 Neuromorphic computing aims at mimicking the synapses, dendrites, and neurons in the brain as well as their associated connected networks to perform a variety of complex tasks including sensing, computing, perception, sometimes by directly utilizing the physical properties of materials. Their functionality diversity and performance highly depend on the use of materials. Compared to the conventional materials, 2D materials exhibit many unique physical properties and the research of 2D materials has reshaped the field of neuromorphic computing. This special issue presents some of the innovations in using devices based on 2D materials to emulate the biological synapses or generate noise injection to hardware neural networks. The issue also provides a comprehensive analysis of recent advances in exploiting the unique physical properties of 2D materials for neuromorphic computing. These innovations and analysis may serve as a useful guide to further advance 2D materials for practical applications. This special issue includes two research articles and four review articles, with contents briefly summarized in the following paragraphs.","PeriodicalId":198030,"journal":{"name":"Neuromorphic Computing and Engineering","volume":"289 1-2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116858854","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：