1
|
Podlaski WF, Machens CK. Approximating Nonlinear Functions With Latent Boundaries in Low-Rank Excitatory-Inhibitory Spiking Networks. Neural Comput 2024; 36:803-857. [PMID: 38658028 DOI: 10.1162/neco_a_01658] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/21/2023] [Accepted: 01/02/2024] [Indexed: 04/26/2024]
Abstract
Deep feedforward and recurrent neural networks have become successful functional models of the brain, but they neglect obvious biological details such as spikes and Dale's law. Here we argue that these details are crucial in order to understand how real neural circuits operate. Towards this aim, we put forth a new framework for spike-based computation in low-rank excitatory-inhibitory spiking networks. By considering populations with rank-1 connectivity, we cast each neuron's spiking threshold as a boundary in a low-dimensional input-output space. We then show how the combined thresholds of a population of inhibitory neurons form a stable boundary in this space, and those of a population of excitatory neurons form an unstable boundary. Combining the two boundaries results in a rank-2 excitatory-inhibitory (EI) network with inhibition-stabilized dynamics at the intersection of the two boundaries. The computation of the resulting networks can be understood as the difference of two convex functions and is thereby capable of approximating arbitrary non-linear input-output mappings. We demonstrate several properties of these networks, including noise suppression and amplification, irregular activity and synaptic balance, as well as how they relate to rate network dynamics in the limit that the boundary becomes soft. Finally, while our work focuses on small networks (5-50 neurons), we discuss potential avenues for scaling up to much larger networks. Overall, our work proposes a new perspective on spiking networks that may serve as a starting point for a mechanistic understanding of biological spike-based computation.
Collapse
Affiliation(s)
- William F Podlaski
- Champalimaud Neuroscience Programme, Champalimaud Foundation, 1400-038 Lisbon, Portugal
| | - Christian K Machens
- Champalimaud Neuroscience Programme, Champalimaud Foundation, 1400-038 Lisbon, Portugal
| |
Collapse
|
2
|
Gouda M, Abreu S, Bienstman P. Surrogate gradient learning in spiking networks trained on event-based cytometry dataset. OPTICS EXPRESS 2024; 32:16260-16272. [PMID: 38859258 DOI: 10.1364/oe.518323] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 01/18/2024] [Accepted: 04/03/2024] [Indexed: 06/12/2024]
Abstract
Spiking neural networks (SNNs) are bio-inspired neural networks that - to an extent - mimic the workings of our brains. In a similar fashion, event-based vision sensors try to replicate a biological eye as closely as possible. In this work, we integrate both technologies for the purpose of classifying micro-particles in the context of label-free flow cytometry. We follow up on our previous work in which we used simple logistic regression with binary labels. Although this model was able to achieve an accuracy of over 98%, our goal is to utilize the system for a wider variety of cells, some of which may have less noticeable morphological variations. Therefore, a more advanced machine learning model like the SNNs discussed here would be required. This comes with the challenge of training such networks, since they typically suffer from vanishing gradients. We effectively apply the surrogate gradient method to overcome this issue achieving over 99% classification accuracy on test data for a four-class problem. Finally, rather than treating the neural network as a black box, we explore the dynamics inside the network and make use of that to enhance its accuracy and sparsity.
Collapse
|
3
|
Deckers L, Van Damme L, Van Leekwijck W, Tsang IJ, Latré S. Co-learning synaptic delays, weights and adaptation in spiking neural networks. Front Neurosci 2024; 18:1360300. [PMID: 38680445 PMCID: PMC11055628 DOI: 10.3389/fnins.2024.1360300] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/22/2023] [Accepted: 03/20/2024] [Indexed: 05/01/2024] Open
Abstract
Spiking neural network (SNN) distinguish themselves from artificial neural network (ANN) because of their inherent temporal processing and spike-based computations, enabling a power-efficient implementation in neuromorphic hardware. In this study, we demonstrate that data processing with spiking neurons can be enhanced by co-learning the synaptic weights with two other biologically inspired neuronal features: (1) a set of parameters describing neuronal adaptation processes and (2) synaptic propagation delays. The former allows a spiking neuron to learn how to specifically react to incoming spikes based on its past. The trained adaptation parameters result in neuronal heterogeneity, which leads to a greater variety in available spike patterns and is also found in the brain. The latter enables to learn to explicitly correlate spike trains that are temporally distanced. Synaptic delays reflect the time an action potential requires to travel from one neuron to another. We show that each of the co-learned features separately leads to an improvement over the baseline SNN and that the combination of both leads to state-of-the-art SNN results on all speech recognition datasets investigated with a simple 2-hidden layer feed-forward network. Our SNN outperforms the benchmark ANN on the neuromorphic datasets (Spiking Heidelberg Digits and Spiking Speech Commands), even with fewer trainable parameters. On the 35-class Google Speech Commands dataset, our SNN also outperforms a GRU of similar size. Our study presents brain-inspired improvements in SNN that enable them to excel over an equivalent ANN of similar size on tasks with rich temporal dynamics.
Collapse
Affiliation(s)
- Lucas Deckers
- IDLab, imec, University of Antwerp, Antwerp, Belgium
| | | | | | | | | |
Collapse
|
4
|
Yan S, Meng Q, Xiao M, Wang Y, Lin Z. Sampling complex topology structures for spiking neural networks. Neural Netw 2024; 172:106121. [PMID: 38244355 DOI: 10.1016/j.neunet.2024.106121] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/11/2023] [Revised: 12/22/2023] [Accepted: 01/09/2024] [Indexed: 01/22/2024]
Abstract
Spiking Neural Networks (SNNs) have been considered a potential competitor to Artificial Neural Networks (ANNs) due to their high biological plausibility and energy efficiency. However, the architecture design of SNN has not been well studied. Previous studies either use ANN architectures or directly search for SNN architectures under a highly constrained search space. In this paper, we aim to introduce much more complex connection topologies to SNNs to further exploit the potential of SNN architectures. To this end, we propose the topology-aware search space, which is the first search space that enables a more diverse and flexible design for both the spatial and temporal topology of the SNN architecture. Then, to efficiently obtain architecture from our search space, we propose the spatio-temporal topology sampling (STTS) algorithm. By leveraging the benefits of random sampling, STTS can yield powerful architecture without the need for an exhaustive search process, making it significantly more efficient than alternative search strategies. Extensive experiments on CIFAR-10, CIFAR-100, and ImageNet demonstrate the effectiveness of our method. Notably, we obtain 70.79% top-1 accuracy on ImageNet with only 4 time steps, 1.79% higher than the second best model. Our code is available under https://github.com/stiger1000/Random-Sampling-SNN.
Collapse
Affiliation(s)
- Shen Yan
- Center for Data Science, Peking University, China.
| | - Qingyan Meng
- The Chinese University of Hong Kong, Shenzhen, China; Shenzhen Research Institute of Big Data, Shenzhen 518115, China.
| | - Mingqing Xiao
- National Key Lab of General AI, School of Intelligence Science and Technology, Peking University, China.
| | - Yisen Wang
- National Key Lab of General AI, School of Intelligence Science and Technology, Peking University, China; Institute for Artificial Intelligence, Peking University, China.
| | - Zhouchen Lin
- National Key Lab of General AI, School of Intelligence Science and Technology, Peking University, China; Institute for Artificial Intelligence, Peking University, China; Peng Cheng Laboratory, Shenzhen, 518055, China.
| |
Collapse
|
5
|
Xu F, Pan D, Zheng H, Ouyang Y, Jia Z, Zeng H. EESCN: A novel spiking neural network method for EEG-based emotion recognition. COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE 2024; 243:107927. [PMID: 38000320 DOI: 10.1016/j.cmpb.2023.107927] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 05/18/2023] [Revised: 10/16/2023] [Accepted: 11/13/2023] [Indexed: 11/26/2023]
Abstract
BACKGROUND AND OBJECTIVE Although existing artificial neural networks have achieved good results in electroencephalograph (EEG) emotion recognition, further improvements are needed in terms of bio-interpretability and robustness. In this research, we aim to develop a highly efficient and high-performance method for emotion recognition based on EEG. METHODS We propose an Emo-EEGSpikeConvNet (EESCN), a novel emotion recognition method based on spiking neural network (SNN). It consists of a neuromorphic data generation module and a NeuroSpiking framework. The neuromorphic data generation module converts EEG data into 2D frame format as input to the NeuroSpiking framework, while the NeuroSpiking framework is used to extract spatio-temporal features of EEG for classification. RESULTS EESCN achieves high emotion recognition accuracies on DEAP and SEED-IV datasets, ranging from 94.56% to 94.81% on DEAP and a mean accuracy of 79.65% on SEED-IV. Compared to existing SNN methods, EESCN significantly improves EEG emotion recognition performance. In addition, it also has the advantages of faster running speed and less memory footprint. CONCLUSIONS EESCN has shown excellent performance and efficiency in EEG-based emotion recognition with potential for practical applications requiring portability and resource constraints.
Collapse
Affiliation(s)
- FeiFan Xu
- Hangzhou Dianzi University, School of Computer Science and Technology, HangZhou, ZheJiang, China.
| | - Deng Pan
- Hangzhou Dianzi University, School of Computer Science and Technology, HangZhou, ZheJiang, China.
| | - Haohao Zheng
- Hangzhou Dianzi University, School of Computer Science and Technology, HangZhou, ZheJiang, China.
| | - Yu Ouyang
- Hangzhou Dianzi University, School of Computer Science and Technology, HangZhou, ZheJiang, China.
| | - Zhe Jia
- Hangzhou Dianzi University, School of Computer Science and Technology, HangZhou, ZheJiang, China.
| | - Hong Zeng
- Hangzhou Dianzi University, School of Computer Science and Technology, HangZhou, ZheJiang, China; Key Laboratory of Brain Machine Collaborative of Zhejiang Province, HangZhou, ZheJiang, China.
| |
Collapse
|
6
|
Sakemi Y, Yamamoto K, Hosomi T, Aihara K. Sparse-firing regularization methods for spiking neural networks with time-to-first-spike coding. Sci Rep 2023; 13:22897. [PMID: 38129555 PMCID: PMC10739753 DOI: 10.1038/s41598-023-50201-5] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/09/2023] [Accepted: 12/16/2023] [Indexed: 12/23/2023] Open
Abstract
The training of multilayer spiking neural networks (SNNs) using the error backpropagation algorithm has made significant progress in recent years. Among the various training schemes, the error backpropagation method that directly uses the firing time of neurons has attracted considerable attention because it can realize ideal temporal coding. This method uses time-to-first-spike (TTFS) coding, in which each neuron fires at most once, and this restriction on the number of firings enables information to be processed at a very low firing frequency. This low firing frequency increases the energy efficiency of information processing in SNNs. However, only an upper limit has been provided for TTFS-coded SNNs, and the information-processing capability of SNNs at lower firing frequencies has not been fully investigated. In this paper, we propose two spike-timing-based sparse-firing (SSR) regularization methods to further reduce the firing frequency of TTFS-coded SNNs. Both methods are characterized by the fact that they only require information about the firing timing and associated weights. The effects of these regularization methods were investigated on the MNIST, Fashion-MNIST, and CIFAR-10 datasets using multilayer perceptron networks and convolutional neural network structures.
Collapse
Affiliation(s)
- Yusuke Sakemi
- Research Center for Mathematical Engineering, Chiba Institute of Technology, Narashino, Japan.
- International Research Center for Neurointelligence (WPI-IRCN), The University of Tokyo, Tokyo, Japan.
| | | | | | - Kazuyuki Aihara
- Research Center for Mathematical Engineering, Chiba Institute of Technology, Narashino, Japan
- International Research Center for Neurointelligence (WPI-IRCN), The University of Tokyo, Tokyo, Japan
| |
Collapse
|
7
|
Schilling A, Sedley W, Gerum R, Metzner C, Tziridis K, Maier A, Schulze H, Zeng FG, Friston KJ, Krauss P. Predictive coding and stochastic resonance as fundamental principles of auditory phantom perception. Brain 2023; 146:4809-4825. [PMID: 37503725 PMCID: PMC10690027 DOI: 10.1093/brain/awad255] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/26/2022] [Revised: 06/27/2023] [Accepted: 07/15/2023] [Indexed: 07/29/2023] Open
Abstract
Mechanistic insight is achieved only when experiments are employed to test formal or computational models. Furthermore, in analogy to lesion studies, phantom perception may serve as a vehicle to understand the fundamental processing principles underlying healthy auditory perception. With a special focus on tinnitus-as the prime example of auditory phantom perception-we review recent work at the intersection of artificial intelligence, psychology and neuroscience. In particular, we discuss why everyone with tinnitus suffers from (at least hidden) hearing loss, but not everyone with hearing loss suffers from tinnitus. We argue that intrinsic neural noise is generated and amplified along the auditory pathway as a compensatory mechanism to restore normal hearing based on adaptive stochastic resonance. The neural noise increase can then be misinterpreted as auditory input and perceived as tinnitus. This mechanism can be formalized in the Bayesian brain framework, where the percept (posterior) assimilates a prior prediction (brain's expectations) and likelihood (bottom-up neural signal). A higher mean and lower variance (i.e. enhanced precision) of the likelihood shifts the posterior, evincing a misinterpretation of sensory evidence, which may be further confounded by plastic changes in the brain that underwrite prior predictions. Hence, two fundamental processing principles provide the most explanatory power for the emergence of auditory phantom perceptions: predictive coding as a top-down and adaptive stochastic resonance as a complementary bottom-up mechanism. We conclude that both principles also play a crucial role in healthy auditory perception. Finally, in the context of neuroscience-inspired artificial intelligence, both processing principles may serve to improve contemporary machine learning techniques.
Collapse
Affiliation(s)
- Achim Schilling
- Neuroscience Lab, University Hospital Erlangen, 91054 Erlangen, Germany
- Cognitive Computational Neuroscience Group, University Erlangen-Nürnberg, 91058 Erlangen, Germany
| | - William Sedley
- Translational and Clinical Research Institute, Newcastle University Medical School, Newcastle upon Tyne NE2 4HH, UK
| | - Richard Gerum
- Cognitive Computational Neuroscience Group, University Erlangen-Nürnberg, 91058 Erlangen, Germany
- Department of Physics and Astronomy and Center for Vision Research, York University, Toronto, ON M3J 1P3, Canada
| | - Claus Metzner
- Neuroscience Lab, University Hospital Erlangen, 91054 Erlangen, Germany
| | | | - Andreas Maier
- Pattern Recognition Lab, University Erlangen-Nürnberg, 91058 Erlangen, Germany
| | - Holger Schulze
- Neuroscience Lab, University Hospital Erlangen, 91054 Erlangen, Germany
| | - Fan-Gang Zeng
- Center for Hearing Research, Departments of Anatomy and Neurobiology, Biomedical Engineering, Cognitive Sciences, Otolaryngology–Head and Neck Surgery, University of California Irvine, Irvine, CA 92697, USA
| | - Karl J Friston
- Wellcome Centre for Human Neuroimaging, Institute of Neurology, University College London, London WC1N 3AR, UK
| | - Patrick Krauss
- Neuroscience Lab, University Hospital Erlangen, 91054 Erlangen, Germany
- Cognitive Computational Neuroscience Group, University Erlangen-Nürnberg, 91058 Erlangen, Germany
- Pattern Recognition Lab, University Erlangen-Nürnberg, 91058 Erlangen, Germany
| |
Collapse
|
8
|
Wu X, Song Y, Zhou Y, Jiang Y, Bai Y, Li X, Yang X. STCA-SNN: self-attention-based temporal-channel joint attention for spiking neural networks. Front Neurosci 2023; 17:1261543. [PMID: 38027490 PMCID: PMC10667472 DOI: 10.3389/fnins.2023.1261543] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/20/2023] [Accepted: 10/23/2023] [Indexed: 12/01/2023] Open
Abstract
Spiking Neural Networks (SNNs) have shown great promise in processing spatio-temporal information compared to Artificial Neural Networks (ANNs). However, there remains a performance gap between SNNs and ANNs, which impedes the practical application of SNNs. With intrinsic event-triggered property and temporal dynamics, SNNs have the potential to effectively extract spatio-temporal features from event streams. To leverage the temporal potential of SNNs, we propose a self-attention-based temporal-channel joint attention SNN (STCA-SNN) with end-to-end training, which infers attention weights along both temporal and channel dimensions concurrently. It models global temporal and channel information correlations with self-attention, enabling the network to learn 'what' and 'when' to attend simultaneously. Our experimental results show that STCA-SNNs achieve better performance on N-MNIST (99.67%), CIFAR10-DVS (81.6%), and N-Caltech 101 (80.88%) compared with the state-of-the-art SNNs. Meanwhile, our ablation study demonstrates that STCA-SNNs improve the accuracy of event stream classification tasks.
Collapse
Affiliation(s)
| | - Yong Song
- School of Optics and Photonics, Beijing Institute of Technology, Beijing, China
| | - Ya Zhou
- School of Optics and Photonics, Beijing Institute of Technology, Beijing, China
| | | | | | | | | |
Collapse
|
9
|
Stanojevic A, Woźniak S, Bellec G, Cherubini G, Pantazi A, Gerstner W. An exact mapping from ReLU networks to spiking neural networks. Neural Netw 2023; 168:74-88. [PMID: 37742533 DOI: 10.1016/j.neunet.2023.09.011] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/24/2023] [Revised: 08/31/2023] [Accepted: 09/04/2023] [Indexed: 09/26/2023]
Abstract
Deep spiking neural networks (SNNs) offer the promise of low-power artificial intelligence. However, training deep SNNs from scratch or converting deep artificial neural networks to SNNs without loss of performance has been a challenge. Here we propose an exact mapping from a network with Rectified Linear Units (ReLUs) to an SNN that fires exactly one spike per neuron. For our constructive proof, we assume that an arbitrary multi-layer ReLU network with or without convolutional layers, batch normalization and max pooling layers was trained to high performance on some training set. Furthermore, we assume that we have access to a representative example of input data used during training and to the exact parameters (weights and biases) of the trained ReLU network. The mapping from deep ReLU networks to SNNs causes zero percent drop in accuracy on CIFAR10, CIFAR100 and the ImageNet-like data sets Places365 and PASS. More generally our work shows that an arbitrary deep ReLU network can be replaced by an energy-efficient single-spike neural network without any loss of performance.
Collapse
Affiliation(s)
- Ana Stanojevic
- IBM Research Europe - Zurich, Rüschlikon, Switzerland; École polytechnique fédérale de Lausanne, School of Life Sciences and School of Computer and Communication Sciences, Lausanne EPFL, Switzerland.
| | | | - Guillaume Bellec
- École polytechnique fédérale de Lausanne, School of Life Sciences and School of Computer and Communication Sciences, Lausanne EPFL, Switzerland
| | | | | | - Wulfram Gerstner
- École polytechnique fédérale de Lausanne, School of Life Sciences and School of Computer and Communication Sciences, Lausanne EPFL, Switzerland
| |
Collapse
|
10
|
Ma G, Yan R, Tang H. Exploiting noise as a resource for computation and learning in spiking neural networks. PATTERNS (NEW YORK, N.Y.) 2023; 4:100831. [PMID: 37876899 PMCID: PMC10591140 DOI: 10.1016/j.patter.2023.100831] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 05/25/2023] [Revised: 07/06/2023] [Accepted: 08/07/2023] [Indexed: 10/26/2023]
Abstract
Networks of spiking neurons underpin the extraordinary information-processing capabilities of the brain and have become pillar models in neuromorphic artificial intelligence. Despite extensive research on spiking neural networks (SNNs), most studies are established on deterministic models, overlooking the inherent non-deterministic, noisy nature of neural computations. This study introduces the noisy SNN (NSNN) and the noise-driven learning (NDL) rule by incorporating noisy neuronal dynamics to exploit the computational advantages of noisy neural processing. The NSNN provides a theoretical framework that yields scalable, flexible, and reliable computation and learning. We demonstrate that this framework leads to spiking neural models with competitive performance, improved robustness against challenging perturbations compared with deterministic SNNs, and better reproducing probabilistic computation in neural coding. Generally, this study offers a powerful and easy-to-use tool for machine learning, neuromorphic intelligence practitioners, and computational neuroscience researchers.
Collapse
Affiliation(s)
- Gehua Ma
- College of Computer Science and Technology, Zhejiang University, Hangzhou, PRC
| | - Rui Yan
- College of Computer Science and Technology, Zhejiang University of Technology, Hangzhou, PRC
| | - Huajin Tang
- College of Computer Science and Technology, Zhejiang University, Hangzhou, PRC
- State Key Lab of Brain-Machine Intelligence, Zhejiang University, Hangzhou, PRC
| |
Collapse
|
11
|
Grimaldi A, Perrinet LU. Learning heterogeneous delays in a layer of spiking neurons for fast motion detection. BIOLOGICAL CYBERNETICS 2023; 117:373-387. [PMID: 37695359 DOI: 10.1007/s00422-023-00975-8] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 09/30/2022] [Accepted: 08/18/2023] [Indexed: 09/12/2023]
Abstract
The precise timing of spikes emitted by neurons plays a crucial role in shaping the response of efferent biological neurons. This temporal dimension of neural activity holds significant importance in understanding information processing in neurobiology, especially for the performance of neuromorphic hardware, such as event-based cameras. Nonetheless, many artificial neural models disregard this critical temporal dimension of neural activity. In this study, we present a model designed to efficiently detect temporal spiking motifs using a layer of spiking neurons equipped with heterogeneous synaptic delays. Our model capitalizes on the diverse synaptic delays present on the dendritic tree, enabling specific arrangements of temporally precise synaptic inputs to synchronize upon reaching the basal dendritic tree. We formalize this process as a time-invariant logistic regression, which can be trained using labeled data. To demonstrate its practical efficacy, we apply the model to naturalistic videos transformed into event streams, simulating the output of the biological retina or event-based cameras. To evaluate the robustness of the model in detecting visual motion, we conduct experiments by selectively pruning weights and demonstrate that the model remains efficient even under significantly reduced workloads. In conclusion, by providing a comprehensive, event-driven computational building block, the incorporation of heterogeneous delays has the potential to greatly improve the performance of future spiking neural network algorithms, particularly in the context of neuromorphic chips.
Collapse
Affiliation(s)
- Antoine Grimaldi
- Institut de Neurosciences de la Timone, Aix Marseille Univ, CNRS, 27 boulevard Jean Moulin, 13005, Marseille, France
| | - Laurent U Perrinet
- Institut de Neurosciences de la Timone, Aix Marseille Univ, CNRS, 27 boulevard Jean Moulin, 13005, Marseille, France.
| |
Collapse
|
12
|
Bitar A, Rosales R, Paulitsch M. Gradient-based feature-attribution explainability methods for spiking neural networks. Front Neurosci 2023; 17:1153999. [PMID: 37829721 PMCID: PMC10565802 DOI: 10.3389/fnins.2023.1153999] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/30/2023] [Accepted: 09/01/2023] [Indexed: 10/14/2023] Open
Abstract
Introduction Spiking neural networks (SNNs) are a model of computation that mimics the behavior of biological neurons. SNNs process event data (spikes) and operate more sparsely than artificial neural networks (ANNs), resulting in ultra-low latency and small power consumption. This paper aims to adapt and evaluate gradient-based explainability methods for SNNs, which were originally developed for conventional ANNs. Methods The adapted methods aim to create input feature attribution maps for SNNs trained through backpropagation that process either event-based spiking data or real-valued data. The methods address the limitations of existing work on explainability methods for SNNs, such as poor scalability, limited to convolutional layers, requiring the training of another model, and providing maps of activation values instead of true attribution scores. The adapted methods are evaluated on classification tasks for both real-valued and spiking data, and the accuracy of the proposed methods is confirmed through perturbation experiments at the pixel and spike levels. Results and discussion The results reveal that gradient-based SNN attribution methods successfully identify highly contributing pixels and spikes with significantly less computation time than model-agnostic methods. Additionally, we observe that the chosen coding technique has a noticeable effect on the input features that will be most significant. These findings demonstrate the potential of gradient-based explainability methods for SNNs in improving our understanding of how these networks process information and contribute to the development of more efficient and accurate SNNs.
Collapse
Affiliation(s)
- Ammar Bitar
- Intel Labs, Munich, Germany
- Department of Knowledge Engineering, Maastricht University, Maastricht, Netherlands
| | | | | |
Collapse
|
13
|
Yao M, Zhang H, Zhao G, Zhang X, Wang D, Cao G, Li G. Sparser spiking activity can be better: Feature Refine-and-Mask spiking neural network for event-based visual recognition. Neural Netw 2023; 166:410-423. [PMID: 37549609 DOI: 10.1016/j.neunet.2023.07.008] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/14/2022] [Revised: 02/23/2023] [Accepted: 07/05/2023] [Indexed: 08/09/2023]
Abstract
Event-based visual, a new visual paradigm with bio-inspired dynamic perception and μs level temporal resolution, has prominent advantages in many specific visual scenarios and gained much research interest. Spiking neural network (SNN) is naturally suitable for dealing with event streams due to its temporal information processing capability and event-driven nature. However, existing works SNN neglect the fact that the input event streams are spatially sparse and temporally non-uniform, and just treat these variant inputs equally. This situation interferes with the effectiveness and efficiency of existing SNNs. In this paper, we propose the feature Refine-and-Mask SNN (RM-SNN), which has the ability of self-adaption to regulate the spiking response in a data-dependent way. We use the Refine-and-Mask (RM) module to refine all features and mask the unimportant features to optimize the membrane potential of spiking neurons, which in turn drops the spiking activity. Inspired by the fact that not all events in spatio-temporal streams are task-relevant, we execute the RM module in both temporal and channel dimensions. Extensive experiments on seven event-based benchmarks, DVS128 Gesture, DVS128 Gait, CIFAR10-DVS, N-Caltech101, DailyAction-DVS, UCF101-DVS, and HMDB51-DVS demonstrate that under the multi-scale constraints of input time window, RM-SNN can significantly reduce the network average spiking activity rate while improving the task performance. In addition, by visualizing spiking responses, we analyze why sparser spiking activity can be better. Code.
Collapse
Affiliation(s)
- Man Yao
- School of Automation Science and Engineering, Xi'an Jiaotong University, Xi'an, Shaanxi 710049, China; Peng Cheng Laboratory, Shenzhen 518000, China.
| | - Hengyu Zhang
- School of Automation Science and Engineering, Xi'an Jiaotong University, Xi'an, Shaanxi 710049, China; Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen 518000, China.
| | - Guangshe Zhao
- School of Automation Science and Engineering, Xi'an Jiaotong University, Xi'an, Shaanxi 710049, China.
| | - Xiyu Zhang
- School of Automation Science and Engineering, Xi'an Jiaotong University, Xi'an, Shaanxi 710049, China.
| | - Dingheng Wang
- Northwest Institute of Mechanical & Electrical Engineering, Xianyang, Shaanxi, China.
| | - Gang Cao
- Beijing Academy of Artificial Intelligence, Beijing 100089, China
| | - Guoqi Li
- Peng Cheng Laboratory, Shenzhen 518000, China; Institute of Automation, Chinese Academy of Sciences, Beijing 100089, China.
| |
Collapse
|
14
|
Mysin I. Phase relations of interneuronal activity relative to theta rhythm. Front Neural Circuits 2023; 17:1198573. [PMID: 37484208 PMCID: PMC10358363 DOI: 10.3389/fncir.2023.1198573] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/01/2023] [Accepted: 06/20/2023] [Indexed: 07/25/2023] Open
Abstract
The theta rhythm plays a crucial role in synchronizing neural activity during attention and memory processes. However, the mechanisms behind the formation of neural activity during theta rhythm generation remain unknown. To address this, we propose a mathematical model that explains the distribution of interneurons in the CA1 field during the theta rhythm phase. Our model consists of a network of seven types of interneurons in the CA1 field that receive inputs from the CA3 field, entorhinal cortex, and local pyramidal neurons in the CA1 field. By adjusting the parameters of the connections in the model. We demonstrate that it is possible to replicate the experimentally observed phase relations between interneurons and the theta rhythm. Our model predicts that populations of interneurons receive unimodal excitation and inhibition with coinciding peaks, and that excitation dominates to determine the firing dynamics of interneurons.
Collapse
|
15
|
Schmidgall S, Hays J. Meta-SpikePropamine: learning to learn with synaptic plasticity in spiking neural networks. Front Neurosci 2023; 17:1183321. [PMID: 37250397 PMCID: PMC10213417 DOI: 10.3389/fnins.2023.1183321] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/09/2023] [Accepted: 04/06/2023] [Indexed: 05/31/2023] Open
Abstract
We propose that in order to harness our understanding of neuroscience toward machine learning, we must first have powerful tools for training brain-like models of learning. Although substantial progress has been made toward understanding the dynamics of learning in the brain, neuroscience-derived models of learning have yet to demonstrate the same performance capabilities as methods in deep learning such as gradient descent. Inspired by the successes of machine learning using gradient descent, we introduce a bi-level optimization framework that seeks to both solve online learning tasks and improve the ability to learn online using models of plasticity from neuroscience. We demonstrate that models of three-factor learning with synaptic plasticity taken from the neuroscience literature can be trained in Spiking Neural Networks (SNNs) with gradient descent via a framework of learning-to-learn to address challenging online learning problems. This framework opens a new path toward developing neuroscience inspired online learning algorithms.
Collapse
Affiliation(s)
- Samuel Schmidgall
- U.S. Naval Research Laboratory, Spacecraft Engineering Department, Washington, DC, United States
- Department of Electrical and Computer Engineering, Johns Hopkins University, Baltimore, MD, United States
| | - Joe Hays
- U.S. Naval Research Laboratory, Spacecraft Engineering Department, Washington, DC, United States
| |
Collapse
|
16
|
Wang D, Wu B, Zhao G, Yao M, Chen H, Deng L, Yan T, Li G. Kronecker CP Decomposition With Fast Multiplication for Compressing RNNs. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2023; 34:2205-2219. [PMID: 34534089 DOI: 10.1109/tnnls.2021.3105961] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 05/04/2023]
Abstract
Recurrent neural networks (RNNs) are powerful in the tasks oriented to sequential data, such as natural language processing and video recognition. However, because the modern RNNs have complex topologies and expensive space/computation complexity, compressing them becomes a hot and promising topic in recent years. Among plenty of compression methods, tensor decomposition, e.g., tensor train (TT), block term (BT), tensor ring (TR), and hierarchical Tucker (HT), appears to be the most amazing approach because a very high compression ratio might be obtained. Nevertheless, none of these tensor decomposition formats can provide both space and computation efficiency. In this article, we consider to compress RNNs based on a novel Kronecker CANDECOMP/PARAFAC (KCP) decomposition, which is derived from Kronecker tensor (KT) decomposition, by proposing two fast algorithms of multiplication between the input and the tensor-decomposed weight. According to our experiments based on UCF11, Youtube Celebrities Face, UCF50, TIMIT, TED-LIUM, and Spiking Heidelberg digits datasets, it can be verified that the proposed KCP-RNNs have a comparable performance of accuracy with those in other tensor-decomposed formats, and even 278 219× compression ratio could be obtained by the low-rank KCP. More importantly, KCP-RNNs are efficient in both space and computation complexity compared with other tensor-decomposed ones. Besides, we find KCP has the best potential of parallel computing to accelerate the calculations in neural networks.
Collapse
|
17
|
Winston CN, Mastrovito D, Shea-Brown E, Mihalas S. Heterogeneity in Neuronal Dynamics Is Learned by Gradient Descent for Temporal Processing Tasks. Neural Comput 2023; 35:555-592. [PMID: 36827598 PMCID: PMC10044000 DOI: 10.1162/neco_a_01571] [Citation(s) in RCA: 2] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/19/2022] [Accepted: 11/02/2022] [Indexed: 02/26/2023]
Abstract
Individual neurons in the brain have complex intrinsic dynamics that are highly diverse. We hypothesize that the complex dynamics produced by networks of complex and heterogeneous neurons may contribute to the brain's ability to process and respond to temporally complex data. To study the role of complex and heterogeneous neuronal dynamics in network computation, we develop a rate-based neuronal model, the generalized-leaky-integrate-and-fire-rate (GLIFR) model, which is a rate equivalent of the generalized-leaky-integrate-and-fire model. The GLIFR model has multiple dynamical mechanisms, which add to the complexity of its activity while maintaining differentiability. We focus on the role of after-spike currents, currents induced or modulated by neuronal spikes, in producing rich temporal dynamics. We use machine learning techniques to learn both synaptic weights and parameters underlying intrinsic dynamics to solve temporal tasks. The GLIFR model allows the use of standard gradient descent techniques rather than surrogate gradient descent, which has been used in spiking neural networks. After establishing the ability to optimize parameters using gradient descent in single neurons, we ask how networks of GLIFR neurons learn and perform on temporally challenging tasks, such as sequential MNIST. We find that these networks learn diverse parameters, which gives rise to diversity in neuronal dynamics, as demonstrated by clustering of neuronal parameters. GLIFR networks have mixed performance when compared to vanilla recurrent neural networks, with higher performance in pixel-by-pixel MNIST but lower in line-by-line MNIST. However, they appear to be more robust to random silencing. We find that the ability to learn heterogeneity and the presence of after-spike currents contribute to these gains in performance. Our work demonstrates both the computational robustness of neuronal complexity and diversity in networks and a feasible method of training such models using exact gradients.
Collapse
Affiliation(s)
- Chloe N Winston
- Departments of Neuroscience and Computer Science, University of Washington, Seattle, WA 98195, U.S.A
- University of Washington Computational Neuroscience Center, Seattle, WA 98195, U.S.A.
| | - Dana Mastrovito
- Allen Institute for Brain Science, Seattle, WA 98109, U.S.A.
| | - Eric Shea-Brown
- University of Washington Computational Neuroscience Center, Seattle, WA 98195, U.S.A
- Allen Institute for Brain Science, Seattle, WA 98109, U.S.A
- Department of Applied Mathematics, University of Washington, Seattle, WA 98195, U.S.A.
| | - Stefan Mihalas
- University of Washington Computational Neuroscience Center, Seattle, WA 98195, U.S.A
- Allen Institute for Brain Science, Seattle, WA 98109, U.S.A
- Department of Applied Mathematics, University of Washington, Seattle, WA 98195, U.S.A.
| |
Collapse
|
18
|
DePasquale B, Sussillo D, Abbott LF, Churchland MM. The centrality of population-level factors to network computation is demonstrated by a versatile approach for training spiking networks. Neuron 2023; 111:631-649.e10. [PMID: 36630961 PMCID: PMC10118067 DOI: 10.1016/j.neuron.2022.12.007] [Citation(s) in RCA: 13] [Impact Index Per Article: 13.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/21/2020] [Revised: 06/17/2022] [Accepted: 12/05/2022] [Indexed: 01/12/2023]
Abstract
Neural activity is often described in terms of population-level factors extracted from the responses of many neurons. Factors provide a lower-dimensional description with the aim of shedding light on network computations. Yet, mechanistically, computations are performed not by continuously valued factors but by interactions among neurons that spike discretely and variably. Models provide a means of bridging these levels of description. We developed a general method for training model networks of spiking neurons by leveraging factors extracted from either data or firing-rate-based networks. In addition to providing a useful model-building framework, this formalism illustrates how reliable and continuously valued factors can arise from seemingly stochastic spiking. Our framework establishes procedures for embedding this property in network models with different levels of realism. The relationship between spikes and factors in such networks provides a foundation for interpreting (and subtly redefining) commonly used quantities such as firing rates.
Collapse
Affiliation(s)
- Brian DePasquale
- Princeton Neuroscience Institute, Princeton University, Princeton NJ, USA; Department of Neuroscience, Columbia University, New York, NY, USA; Center for Theoretical Neuroscience, Columbia University, New York, NY, USA.
| | - David Sussillo
- Department of Electrical Engineering, Stanford University, Stanford, CA, USA; Wu Tsai Neurosciences Institute, Stanford University, Stanford, CA, USA
| | - L F Abbott
- Department of Neuroscience, Columbia University, New York, NY, USA; Center for Theoretical Neuroscience, Columbia University, New York, NY, USA; Zuckerman Mind Brain Behavior Institute, Columbia University, New York, NY, USA; Department of Physiology and Cellular Biophysics, Columbia University, New York, NY, USA; Kavli Institute for Brain Science, Columbia University, New York, NY, USA
| | - Mark M Churchland
- Department of Neuroscience, Columbia University, New York, NY, USA; Zuckerman Mind Brain Behavior Institute, Columbia University, New York, NY, USA; Kavli Institute for Brain Science, Columbia University, New York, NY, USA; Grossman Center for the Statistics of Mind, Columbia University, New York, NY, USA
| |
Collapse
|
19
|
Akl M, Ergene D, Walter F, Knoll A. Toward robust and scalable deep spiking reinforcement learning. Front Neurorobot 2023; 16:1075647. [PMID: 36742191 PMCID: PMC9894879 DOI: 10.3389/fnbot.2022.1075647] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/20/2022] [Accepted: 12/23/2022] [Indexed: 01/21/2023] Open
Abstract
Deep reinforcement learning (DRL) combines reinforcement learning algorithms with deep neural networks (DNNs). Spiking neural networks (SNNs) have been shown to be a biologically plausible and energy efficient alternative to DNNs. Since the introduction of surrogate gradient approaches that allowed to overcome the discontinuity in the spike function, SNNs can now be trained with the backpropagation through time (BPTT) algorithm. While largely explored on supervised learning problems, little work has been done on investigating the use of SNNs as function approximators in DRL. Here we show how SNNs can be applied to different DRL algorithms like Deep Q-Network (DQN) and Twin-Delayed Deep Deteministic Policy Gradient (TD3) for discrete and continuous action space environments, respectively. We found that SNNs are sensitive to the additional hyperparameters introduced by spiking neuron models like current and voltage decay factors, firing thresholds, and that extensive hyperparameter tuning is inevitable. However, we show that increasing the simulation time of SNNs, as well as applying a two-neuron encoding to the input observations helps reduce the sensitivity to the membrane parameters. Furthermore, we show that randomizing the membrane parameters, instead of selecting uniform values for all neurons, has stabilizing effects on the training. We conclude that SNNs can be utilized for learning complex continuous control problems with state-of-the-art DRL algorithms. While the training complexity increases, the resulting SNNs can be directly executed on neuromorphic processors and potentially benefit from their high energy efficiency.
Collapse
|
20
|
Precise Spiking Motifs in Neurobiological and Neuromorphic Data. Brain Sci 2022; 13:brainsci13010068. [PMID: 36672049 PMCID: PMC9856822 DOI: 10.3390/brainsci13010068] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/16/2022] [Revised: 12/20/2022] [Accepted: 12/23/2022] [Indexed: 12/31/2022] Open
Abstract
Why do neurons communicate through spikes? By definition, spikes are all-or-none neural events which occur at continuous times. In other words, spikes are on one side binary, existing or not without further details, and on the other, can occur at any asynchronous time, without the need for a centralized clock. This stands in stark contrast to the analog representation of values and the discretized timing classically used in digital processing and at the base of modern-day neural networks. As neural systems almost systematically use this so-called event-based representation in the living world, a better understanding of this phenomenon remains a fundamental challenge in neurobiology in order to better interpret the profusion of recorded data. With the growing need for intelligent embedded systems, it also emerges as a new computing paradigm to enable the efficient operation of a new class of sensors and event-based computers, called neuromorphic, which could enable significant gains in computation time and energy consumption-a major societal issue in the era of the digital economy and global warming. In this review paper, we provide evidence from biology, theory and engineering that the precise timing of spikes plays a crucial role in our understanding of the efficiency of neural networks.
Collapse
|
21
|
Müller-Cleve SF, Fra V, Khacef L, Pequeño-Zurro A, Klepatsch D, Forno E, Ivanovich DG, Rastogi S, Urgese G, Zenke F, Bartolozzi C. Braille letter reading: A benchmark for spatio-temporal pattern recognition on neuromorphic hardware. Front Neurosci 2022; 16:951164. [PMID: 36440280 PMCID: PMC9695069 DOI: 10.3389/fnins.2022.951164] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/23/2022] [Accepted: 10/19/2022] [Indexed: 03/25/2024] Open
Abstract
Spatio-temporal pattern recognition is a fundamental ability of the brain which is required for numerous real-world activities. Recent deep learning approaches have reached outstanding accuracies in such tasks, but their implementation on conventional embedded solutions is still very computationally and energy expensive. Tactile sensing in robotic applications is a representative example where real-time processing and energy efficiency are required. Following a brain-inspired computing approach, we propose a new benchmark for spatio-temporal tactile pattern recognition at the edge through Braille letter reading. We recorded a new Braille letters dataset based on the capacitive tactile sensors of the iCub robot's fingertip. We then investigated the importance of spatial and temporal information as well as the impact of event-based encoding on spike-based computation. Afterward, we trained and compared feedforward and recurrent Spiking Neural Networks (SNNs) offline using Backpropagation Through Time (BPTT) with surrogate gradients, then we deployed them on the Intel Loihi neuromorphic chip for fast and efficient inference. We compared our approach to standard classifiers, in particular to the Long Short-Term Memory (LSTM) deployed on the embedded NVIDIA Jetson GPU, in terms of classification accuracy, power, and energy consumption together with computational delay. Our results show that the LSTM reaches ~97% of accuracy, outperforming the recurrent SNN by ~17% when using continuous frame-based data instead of event-based inputs. However, the recurrent SNN on Loihi with event-based inputs is ~500 times more energy-efficient than the LSTM on Jetson, requiring a total power of only ~30 mW. This work proposes a new benchmark for tactile sensing and highlights the challenges and opportunities of event-based encoding, neuromorphic hardware, and spike-based computing for spatio-temporal pattern recognition at the edge.
Collapse
Affiliation(s)
| | - Vittorio Fra
- Politecnico di Torino, Electronic Design Automation (EDA) Group, Torino, Italy
| | - Lyes Khacef
- Bio-Inspired Circuits and Systems Lab, Zernike Institute for Advanced Materials, Groningen Cognitive Systems and Materials Center, University of Groningen, Groningen, Netherlands
| | | | - Daniel Klepatsch
- Silicon Austria Labs, Johannes Kepler Universität (JKU) Linz Institute of Technology (LIT) Silicon Austria Labs (SAL) embedded Signal Processing and Machine Learning (eSPML) Lab, Graz, Austria
- Johannes Kepler Universität (JKU) Linz Institute of Technology (LIT) Silicon Austria Labs (SAL) embedded Signal Processing and Machine Learning (eSPML) Lab, Johannes Kepler University Linz, Graz, Austria
| | - Evelina Forno
- Politecnico di Torino, Electronic Design Automation (EDA) Group, Torino, Italy
| | - Diego G. Ivanovich
- Silicon Austria Labs, Johannes Kepler Universität (JKU) Linz Institute of Technology (LIT) Silicon Austria Labs (SAL) embedded Signal Processing and Machine Learning (eSPML) Lab, Graz, Austria
- Johannes Kepler Universität (JKU) Linz Institute of Technology (LIT) Silicon Austria Labs (SAL) embedded Signal Processing and Machine Learning (eSPML) Lab, Johannes Kepler University Linz, Graz, Austria
| | - Shavika Rastogi
- International Centre for Neuromorphic Systems, Western Sydney University, Penrith, NSW, Australia
- Biocomputation Research Group, University of Hertfordshire, Hatfield, United Kingdom
| | - Gianvito Urgese
- Politecnico di Torino, Electronic Design Automation (EDA) Group, Torino, Italy
| | - Friedemann Zenke
- Friedrich Miescher Institute for Biomedical Research, Basel, Switzerland
| | - Chiara Bartolozzi
- Istituto Italiano di Tecnologia, Event-Driven Perception in Robotics, Genoa, Italy
| |
Collapse
|
22
|
Yu C, Du Y, Chen M, Wang A, Wang G, Li E. MAP-SNN: Mapping spike activities with multiplicity, adaptability, and plasticity into bio-plausible spiking neural networks. Front Neurosci 2022; 16:945037. [PMID: 36203801 PMCID: PMC9531034 DOI: 10.3389/fnins.2022.945037] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/16/2022] [Accepted: 08/29/2022] [Indexed: 11/26/2022] Open
Abstract
Spiking Neural Networks (SNNs) are considered more biologically realistic and power-efficient as they imitate the fundamental mechanism of the human brain. Backpropagation (BP) based SNN learning algorithms that utilize deep learning frameworks have achieved good performance. However, those BP-based algorithms partially ignore bio-interpretability. In modeling spike activity for biological plausible BP-based SNNs, we examine three properties: multiplicity, adaptability, and plasticity (MAP). Regarding multiplicity, we propose a Multiple-Spike Pattern (MSP) with multiple-spike transmission to improve model robustness in discrete time iterations. To realize adaptability, we adopt Spike Frequency Adaption (SFA) under MSP to reduce spike activities for enhanced efficiency. For plasticity, we propose a trainable state-free synapse that models spike response current to increase the diversity of spiking neurons for temporal feature extraction. The proposed SNN model achieves competitive performances on the N-MNIST and SHD neuromorphic datasets. In addition, experimental results demonstrate that the proposed three aspects are significant to iterative robustness, spike efficiency, and the capacity to extract spikes' temporal features. In summary, this study presents a realistic approach for bio-inspired spike activity with MAP, presenting a novel neuromorphic perspective for incorporating biological properties into spiking neural networks.
Collapse
Affiliation(s)
- Chengting Yu
- College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou, China
- Zhejiang University - University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, China
| | - Yangkai Du
- College of Computer Science and Technology, Zhejiang University, Hangzhou, China
| | - Mufeng Chen
- College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou, China
| | - Aili Wang
- College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou, China
- Zhejiang University - University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, China
- *Correspondence: Aili Wang
| | - Gaoang Wang
- Zhejiang University - University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, China
| | - Erping Li
- College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou, China
- Zhejiang University - University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, China
| |
Collapse
|
23
|
Relaxation LIF: A gradient-based spiking neuron for direct training deep spiking neural networks. Neurocomputing 2022. [DOI: 10.1016/j.neucom.2022.06.036] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/21/2022]
|
24
|
Makarov VA, Lobov SA, Shchanikov S, Mikhaylov A, Kazantsev VB. Toward Reflective Spiking Neural Networks Exploiting Memristive Devices. Front Comput Neurosci 2022; 16:859874. [PMID: 35782090 PMCID: PMC9243340 DOI: 10.3389/fncom.2022.859874] [Citation(s) in RCA: 5] [Impact Index Per Article: 2.5] [Reference Citation Analysis] [Abstract] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/21/2022] [Accepted: 05/10/2022] [Indexed: 11/29/2022] Open
Abstract
The design of modern convolutional artificial neural networks (ANNs) composed of formal neurons copies the architecture of the visual cortex. Signals proceed through a hierarchy, where receptive fields become increasingly more complex and coding sparse. Nowadays, ANNs outperform humans in controlled pattern recognition tasks yet remain far behind in cognition. In part, it happens due to limited knowledge about the higher echelons of the brain hierarchy, where neurons actively generate predictions about what will happen next, i.e., the information processing jumps from reflex to reflection. In this study, we forecast that spiking neural networks (SNNs) can achieve the next qualitative leap. Reflective SNNs may take advantage of their intrinsic dynamics and mimic complex, not reflex-based, brain actions. They also enable a significant reduction in energy consumption. However, the training of SNNs is a challenging problem, strongly limiting their deployment. We then briefly overview new insights provided by the concept of a high-dimensional brain, which has been put forward to explain the potential power of single neurons in higher brain stations and deep SNN layers. Finally, we discuss the prospect of implementing neural networks in memristive systems. Such systems can densely pack on a chip 2D or 3D arrays of plastic synaptic contacts directly processing analog information. Thus, memristive devices are a good candidate for implementing in-memory and in-sensor computing. Then, memristive SNNs can diverge from the development of ANNs and build their niche, cognitive, or reflective computations.
Collapse
Affiliation(s)
- Valeri A. Makarov
- Instituto de Matemática Interdisciplinar, Universidad Complutense de Madrid, Madrid, Spain
- Department of Neurotechnologies, Research Institute of Physics and Technology, Laboratory of Stochastic Multistable Systems, Lobachevsky State University of Nizhny Novgorod, Nizhny Novgorod, Russia
- *Correspondence: Valeri A. Makarov,
| | - Sergey A. Lobov
- Department of Neurotechnologies, Research Institute of Physics and Technology, Laboratory of Stochastic Multistable Systems, Lobachevsky State University of Nizhny Novgorod, Nizhny Novgorod, Russia
- Neuroscience and Cognitive Technology Laboratory, Center for Technologies in Robotics and Mechatronics Components, Innopolis University, Innopolis, Russia
- Center For Neurotechnology and Machine Learning, Immanuel Kant Baltic Federal University, Kaliningrad, Russia
| | - Sergey Shchanikov
- Department of Neurotechnologies, Research Institute of Physics and Technology, Laboratory of Stochastic Multistable Systems, Lobachevsky State University of Nizhny Novgorod, Nizhny Novgorod, Russia
- Department of Information Technologies, Vladimir State University, Vladimir, Russia
| | - Alexey Mikhaylov
- Department of Neurotechnologies, Research Institute of Physics and Technology, Laboratory of Stochastic Multistable Systems, Lobachevsky State University of Nizhny Novgorod, Nizhny Novgorod, Russia
| | - Viktor B. Kazantsev
- Department of Neurotechnologies, Research Institute of Physics and Technology, Laboratory of Stochastic Multistable Systems, Lobachevsky State University of Nizhny Novgorod, Nizhny Novgorod, Russia
- Neuroscience and Cognitive Technology Laboratory, Center for Technologies in Robotics and Mechatronics Components, Innopolis University, Innopolis, Russia
- Center For Neurotechnology and Machine Learning, Immanuel Kant Baltic Federal University, Kaliningrad, Russia
| |
Collapse
|
25
|
Rao A, Plank P, Wild A, Maass W. A Long Short-Term Memory for AI Applications in Spike-based Neuromorphic Hardware. NAT MACH INTELL 2022. [DOI: 10.1038/s42256-022-00480-w] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/09/2022]
|
26
|
Lu S, Sengupta A. Neuroevolution Guided Hybrid Spiking Neural Network Training. Front Neurosci 2022; 16:838523. [PMID: 35546880 PMCID: PMC9082355 DOI: 10.3389/fnins.2022.838523] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/17/2021] [Accepted: 03/11/2022] [Indexed: 11/16/2022] Open
Abstract
Neuromorphic computing algorithms based on Spiking Neural Networks (SNNs) are evolving to be a disruptive technology driving machine learning research. The overarching goal of this work is to develop a structured algorithmic framework for SNN training that optimizes unique SNN-specific properties like neuron spiking threshold using neuroevolution as a feedback strategy. We provide extensive results for this hybrid bio-inspired training strategy and show that such a feedback-based learning approach leads to explainable neuromorphic systems that adapt to the specific underlying application. Our analysis reveals 53.8, 28.8, and 28.2% latency improvement for the neuroevolution-based SNN training strategy on CIFAR-10, CIFAR-100, and ImageNet datasets, respectively in contrast to state-of-the-art conversion based approaches. The proposed algorithm can be easily extended to other application domains like image classification in presence of adversarial attacks where 43.2 and 27.9% latency improvements were observed on CIFAR-10 and CIFAR-100 datasets, respectively.
Collapse
Affiliation(s)
- Sen Lu
- School of Electrical Engineering and Computer Science, The Pennsylvania State University, University Park, PA, United States
| | - Abhronil Sengupta
- School of Electrical Engineering and Computer Science, The Pennsylvania State University, University Park, PA, United States
| |
Collapse
|
27
|
Zou Z, Alimohamadi H, Zakeri A, Imani F, Kim Y, Najafi MH, Imani M. Memory-inspired spiking hyperdimensional network for robust online learning. Sci Rep 2022; 12:7641. [PMID: 35538126 PMCID: PMC9090930 DOI: 10.1038/s41598-022-11073-3] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/14/2021] [Accepted: 04/08/2022] [Indexed: 11/09/2022] Open
Abstract
Recently, brain-inspired computing models have shown great potential to outperform today's deep learning solutions in terms of robustness and energy efficiency. Particularly, Spiking Neural Networks (SNNs) and HyperDimensional Computing (HDC) have shown promising results in enabling efficient and robust cognitive learning. Despite the success, these two brain-inspired models have different strengths. While SNN mimics the physical properties of the human brain, HDC models the brain on a more abstract and functional level. Their design philosophies demonstrate complementary patterns that motivate their combination. With the help of the classical psychological model on memory, we propose SpikeHD, the first framework that fundamentally combines Spiking neural network and hyperdimensional computing. SpikeHD generates a scalable and strong cognitive learning system that better mimics brain functionality. SpikeHD exploits spiking neural networks to extract low-level features by preserving the spatial and temporal correlation of raw event-based spike data. Then, it utilizes HDC to operate over SNN output by mapping the signal into high-dimensional space, learning the abstract information, and classifying the data. Our extensive evaluation on a set of benchmark classification problems shows that SpikeHD provides the following benefit compared to SNN architecture: (1) significantly enhance learning capability by exploiting two-stage information processing, (2) enables substantial robustness to noise and failure, and (3) reduces the network size and required parameters to learn complex information.
Collapse
Affiliation(s)
- Zhuowen Zou
- University of California San Diego, La Jolla, CA, 92093, USA
- University of California Irvine, Irvine, CA, 92697, USA
| | | | - Ali Zakeri
- University of California Irvine, Irvine, CA, 92697, USA
| | - Farhad Imani
- University of Connecticut, Storrs, CT, 06269, USA
| | - Yeseong Kim
- Daegu Gyeongbuk Institute of Science and Technology, Daegu, South Korea
| | | | - Mohsen Imani
- University of California Irvine, Irvine, CA, 92697, USA.
| |
Collapse
|
28
|
Keijser J, Sprekeler H. Optimizing interneuron circuits for compartment-specific feedback inhibition. PLoS Comput Biol 2022; 18:e1009933. [PMID: 35482670 PMCID: PMC9049365 DOI: 10.1371/journal.pcbi.1009933] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/24/2021] [Accepted: 02/18/2022] [Indexed: 12/02/2022] Open
Abstract
Cortical circuits process information by rich recurrent interactions between excitatory neurons and inhibitory interneurons. One of the prime functions of interneurons is to stabilize the circuit by feedback inhibition, but the level of specificity on which inhibitory feedback operates is not fully resolved. We hypothesized that inhibitory circuits could enable separate feedback control loops for different synaptic input streams, by means of specific feedback inhibition to different neuronal compartments. To investigate this hypothesis, we adopted an optimization approach. Leveraging recent advances in training spiking network models, we optimized the connectivity and short-term plasticity of interneuron circuits for compartment-specific feedback inhibition onto pyramidal neurons. Over the course of the optimization, the interneurons diversified into two classes that resembled parvalbumin (PV) and somatostatin (SST) expressing interneurons. Using simulations and mathematical analyses, we show that the resulting circuit can be understood as a neural decoder that inverts the nonlinear biophysical computations performed within the pyramidal cells. Our model provides a proof of concept for studying structure-function relations in cortical circuits by a combination of gradient-based optimization and biologically plausible phenomenological models. The brain contains billions of nerve cells—neurons—that can be classified into different types depending on their shape, connectivity and activity. A particularly diverse group of neurons is that of inhibitory neurons, named after their suppressive effect on neural activity. Presumably, their diverse properties allow inhibitory neurons to fulfil different functions, but what these functions are is currently unclear. In this paper, we investigated if a particular function can explain the existence and properties of the two most common inhibitory cell classes: The need to regulate activity in different physical parts (compartments) of the neurons they target. We investigated this function in a computer model, using optimization to find the neuron properties best-suited for compartment-specific inhibition. Our key result is that after the optimization, model neurons largely fell into two classes that resembled the two types of biological neurons. In particular, the optimized neurons were connected to only one compartment of other neurons. This suggests that the diversity of inhibitory neurons is well suited for compartment-specific inhibition. In the future, our approach of optimizing neural properties might be used to investigate other functions (or dysfunctions) of neuron diversity.
Collapse
Affiliation(s)
- Joram Keijser
- Modelling of Cognitive Processes, Institute of Software Engineering and Theoretical Computer Science, Technische Universität Berlin, Berlin, Germany
- Charité – Universitätsmedizin Berlin, Einstein Center for Neurosciences Berlin, Berlin, Germany
- * E-mail:
| | - Henning Sprekeler
- Modelling of Cognitive Processes, Institute of Software Engineering and Theoretical Computer Science, Technische Universität Berlin, Berlin, Germany
- Bernstein Center for Computational Neuroscience Berlin, Berlin, Germany
| |
Collapse
|
29
|
Analysis of Ice and Snow Path Planning System Based on MNN Algorithm. COMPUTATIONAL INTELLIGENCE AND NEUROSCIENCE 2022; 2022:1586006. [PMID: 35295272 PMCID: PMC8920659 DOI: 10.1155/2022/1586006] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 12/23/2021] [Accepted: 01/20/2022] [Indexed: 11/17/2022]
Abstract
Traditional ice and snow path planning methods still have internal environmental problems in intelligent path planning, such as weak innovation ability, imperfect management, long planning path, unreasonable security structure, and low degree of specialization. Therefore, more and more ice and snow sports lovers are eager to solve this problem. This paper designs a path planning method based on three-dimensional ice and snow model. The path planning method of moving snow and ice based on MNN (Multiclass Neural Networks) algorithm is studied from many aspects. MNN algorithm is used for comprehensive analysis and evaluation. The mobile phone provides data information on key nodes, air resistance, momentum change, ice and snow movement track, and so on. The results show that the ice and snow path planning system based on MNN algorithm designed in this paper has the advantages of high feasibility, high data accuracy, and good prediction effect and can effectively improve the efficiency of ice and snow path planning.
Collapse
|
30
|
Wu Z, Zhang Z, Gao H, Qin J, Zhao R, Zhao G, Li G. Modeling learnable electrical synapse for high precision spatio-temporal recognition. Neural Netw 2022; 149:184-194. [DOI: 10.1016/j.neunet.2022.02.006] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/12/2021] [Revised: 11/30/2021] [Accepted: 02/06/2022] [Indexed: 10/19/2022]
|
31
|
Schirner M, Kong X, Yeo BTT, Deco G, Ritter P. Dynamic primitives of brain network interaction Special Issue "Advances in Mapping the Connectome". Neuroimage 2022; 250:118928. [PMID: 35101596 DOI: 10.1016/j.neuroimage.2022.118928] [Citation(s) in RCA: 11] [Impact Index Per Article: 5.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/01/2021] [Revised: 12/03/2021] [Accepted: 01/20/2022] [Indexed: 01/04/2023] Open
Abstract
What dynamic processes underly functional brain networks? Functional connectivity (FC) and functional connectivity dynamics (FCD) are used to represent the patterns and dynamics of functional brain networks. FC(D) is related to the synchrony of brain activity: when brain areas oscillate in a coordinated manner this yields a high correlation between their signal time series. To explain the processes underlying FC(D) we review how synchronized oscillations emerge from coupled neural populations in brain network models (BNMs). From detailed spiking networks to more abstract population models, there is strong support for the idea that the brain operates near critical instabilities that give rise to multistable or metastable dynamics that in turn lead to the intermittently synchronized slow oscillations underlying FC(D). We explore further consequences from these fundamental mechanisms and how they fit with reality. We conclude by highlighting the need for integrative brain models that connect separate mechanisms across levels of description and spatiotemporal scales and link them with cognitive function.
Collapse
Affiliation(s)
- Michael Schirner
- Berlin Institute of Health at Charité - Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin, Germany; Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt Universität zu Berlin, Department of Neurology with Experimental Neurology, Charitéplatz 1, 10117 Berlin, Germany; Bernstein Focus State Dependencies of Learning & Bernstein Center for Computational Neuroscience, Berlin, Germany; Einstein Center for Neuroscience Berlin, Charitéplatz 1, 10117 Berlin, Germany; Einstein Center Digital Future, Wilhelmstraße 67, 10117 Berlin, Germany.
| | - Xiaolu Kong
- Department of Electrical and Computer Engineering, National University of Singapore, Singapore; Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, Yong Loo Lin School of Medicine, Singapore; N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore, Singapore
| | - B T Thomas Yeo
- Department of Electrical and Computer Engineering, National University of Singapore, Singapore; Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, Yong Loo Lin School of Medicine, Singapore; N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore, Singapore; Integrative Sciences and Engineering Programme (ISEP), National University of Singapore, Singapore, Singapore; Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, USA
| | - Gustavo Deco
- Center for Brain and Cognition, Computational Neuroscience Group, Department of Information and Communication Technologies, Universitat Pompeu Fabra, Barcelona, Spain; Institució Catalana de la Recerca i Estudis Avançats, Barcelona, Spain; Department of Neuropsychology, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany; School of Psychological Sciences, Turner Institute for Brain and Mental Health, Monash University, Melbourne, Clayton, Australia
| | - Petra Ritter
- Berlin Institute of Health at Charité - Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin, Germany; Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt Universität zu Berlin, Department of Neurology with Experimental Neurology, Charitéplatz 1, 10117 Berlin, Germany; Bernstein Focus State Dependencies of Learning & Bernstein Center for Computational Neuroscience, Berlin, Germany; Einstein Center for Neuroscience Berlin, Charitéplatz 1, 10117 Berlin, Germany; Einstein Center Digital Future, Wilhelmstraße 67, 10117 Berlin, Germany.
| |
Collapse
|
32
|
Abstract
Neuromorphic systems aim to accomplish efficient computation in electronics by mirroring neurobiological principles. Taking advantage of neuromorphic technologies requires effective learning algorithms capable of instantiating high-performing neural networks, while also dealing with inevitable manufacturing variations of individual components, such as memristors or analog neurons. We present a learning framework resulting in bioinspired spiking neural networks with high performance, low inference latency, and sparse spike-coding schemes, which also self-corrects for device mismatch. We validate our approach on the BrainScaleS-2 analog spiking neuromorphic system, demonstrating state-of-the-art accuracy, low latency, and energy efficiency. Our work sketches a path for building powerful neuromorphic processors that take advantage of emerging analog technologies. To rapidly process temporal information at a low metabolic cost, biological neurons integrate inputs as an analog sum, but communicate with spikes, binary events in time. Analog neuromorphic hardware uses the same principles to emulate spiking neural networks with exceptional energy efficiency. However, instantiating high-performing spiking networks on such hardware remains a significant challenge due to device mismatch and the lack of efficient training algorithms. Surrogate gradient learning has emerged as a promising training strategy for spiking networks, but its applicability for analog neuromorphic systems has not been demonstrated. Here, we demonstrate surrogate gradient learning on the BrainScaleS-2 analog neuromorphic system using an in-the-loop approach. We show that learning self-corrects for device mismatch, resulting in competitive spiking network performance on both vision and speech benchmarks. Our networks display sparse spiking activity with, on average, less than one spike per hidden neuron and input, perform inference at rates of up to 85,000 frames per second, and consume less than 200 mW. In summary, our work sets several benchmarks for low-energy spiking network processing on analog neuromorphic hardware and paves the way for future on-chip learning algorithms.
Collapse
|
33
|
Eshraghian JK, Wang X, Lu WD. Memristor-Based Binarized Spiking Neural Networks: Challenges and Applications. IEEE NANOTECHNOLOGY MAGAZINE 2022. [DOI: 10.1109/mnano.2022.3141443] [Citation(s) in RCA: 5] [Impact Index Per Article: 2.5] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/10/2022]
|
34
|
Optimization of Data Mining and Analysis System for Chinese Language Teaching Based on Convolutional Neural Network. COMPUTATIONAL INTELLIGENCE AND NEUROSCIENCE 2021; 2021:1148954. [PMID: 34899886 PMCID: PMC8664500 DOI: 10.1155/2021/1148954] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 08/19/2021] [Revised: 10/29/2021] [Accepted: 11/05/2021] [Indexed: 11/17/2022]
Abstract
Chinese language is also an important way to understand Chinese culture and an important carrier to inherit and carry forward Chinese traditional culture. Chinese language teaching is an important way to inherit and develop Chinese language. Therefore, in the era of big data, data mining and analysis of Chinese language teaching can effectively sum up experience and draw lessons, so as to improve the quality of Chinese language teaching and promote Chinese language culture. Text clustering technology can analyze and process the text information data and divide the text information data with the same characteristics into the same category. Based on big data, combined with convolutional neural network and K-means algorithm, this paper proposes a text clustering method based on convolutional neural network (CNN), constructs a Chinese language teaching data mining analysis system, and optimizes it so that the system can better mine Chinese character data in Chinese language teaching data in depth and comprehensively. The results show that the optimized k-means algorithm needs 683 iterations to achieve the target accuracy. The average K-measure value of the optimized system is 0.770, which is higher than that of the original system. The results also show that K-means algorithm can significantly improve the clustering effect, optimize the data mining analysis system of Chinese language teaching, and deeply mine the Chinese data in Chinese language teaching, so as to improve the quality of Chinese language teaching.
Collapse
|
35
|
Spiking Neural Networks for Computational Intelligence: An Overview. BIG DATA AND COGNITIVE COMPUTING 2021. [DOI: 10.3390/bdcc5040067] [Citation(s) in RCA: 5] [Impact Index Per Article: 1.7] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 11/16/2022]
Abstract
Deep neural networks with rate-based neurons have exhibited tremendous progress in the last decade. However, the same level of progress has not been observed in research on spiking neural networks (SNN), despite their capability to handle temporal data, energy-efficiency and low latency. This could be because the benchmarking techniques for SNNs are based on the methods used for evaluating deep neural networks, which do not provide a clear evaluation of the capabilities of SNNs. Particularly, the benchmarking of SNN approaches with regards to energy efficiency and latency requires realization in suitable hardware, which imposes additional temporal and resource constraints upon ongoing projects. This review aims to provide an overview of the current real-world applications of SNNs and identifies steps to accelerate research involving SNNs in the future.
Collapse
|
36
|
Li D. Human Skeleton Detection and Extraction in Dance Video Based on PSO-Enabled LSTM Neural Network. COMPUTATIONAL INTELLIGENCE AND NEUROSCIENCE 2021; 2021:2545151. [PMID: 34552625 PMCID: PMC8452444 DOI: 10.1155/2021/2545151] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 07/19/2021] [Revised: 08/30/2021] [Accepted: 09/01/2021] [Indexed: 11/30/2022]
Abstract
With the significant increase of social informatization, the emerging information technology represented by machine vision has been applied to more and more scenes. Among them, the detection and extraction of human skeleton in a dance video based on this technology has a huge market demand in education and training. However, the existing detection and extraction technology has the problems of slow recognition speed and low extraction accuracy. Therefore, this paper proposes a neural network based on particle swarm optimization to detect and extract human skeletons in a dance video. Through the research and test on different data sets, it is found that the neural network based on particle swarm optimization algorithm has good detection and extraction ability and has high accuracy for the detection and recognition of human skeleton points. Among them, on all MPII data sets, the average accuracy of PSO-LSTM proposed in this paper is 3.9% higher than that of other optimal algorithms; on the PoseTrack data set, the average accuracy of detection and extraction is improved by 2.3%. The above results show that the neural network based on particle swarm optimization has fast detection speed and good extraction accuracy and can be used for the detection and extraction of human skeleton in a dance video.
Collapse
Affiliation(s)
- Dingxin Li
- Department of Sports and Public Art, Zhengzhou University of Aeronautics, Zhengzhou, Henan 450046, China
| |
Collapse
|
37
|
Salaj D, Subramoney A, Kraisnikovic C, Bellec G, Legenstein R, Maass W. Spike frequency adaptation supports network computations on temporally dispersed information. eLife 2021; 10:e65459. [PMID: 34310281 PMCID: PMC8313230 DOI: 10.7554/elife.65459] [Citation(s) in RCA: 15] [Impact Index Per Article: 5.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/04/2020] [Accepted: 06/29/2021] [Indexed: 11/13/2022] Open
Abstract
For solving tasks such as recognizing a song, answering a question, or inverting a sequence of symbols, cortical microcircuits need to integrate and manipulate information that was dispersed over time during the preceding seconds. Creating biologically realistic models for the underlying computations, especially with spiking neurons and for behaviorally relevant integration time spans, is notoriously difficult. We examine the role of spike frequency adaptation in such computations and find that it has a surprisingly large impact. The inclusion of this well-known property of a substantial fraction of neurons in the neocortex - especially in higher areas of the human neocortex - moves the performance of spiking neural network models for computations on network inputs that are temporally dispersed from a fairly low level up to the performance level of the human brain.
Collapse
Affiliation(s)
- Darjan Salaj
- Institute of Theoretical Computer Science, Graz University of TechnologyGrazAustria
| | - Anand Subramoney
- Institute of Theoretical Computer Science, Graz University of TechnologyGrazAustria
| | - Ceca Kraisnikovic
- Institute of Theoretical Computer Science, Graz University of TechnologyGrazAustria
| | - Guillaume Bellec
- Institute of Theoretical Computer Science, Graz University of TechnologyGrazAustria
- Laboratory of Computational Neuroscience, Ecole Polytechnique Fédérale de Lausanne (EPFL)LausanneSwitzerland
| | - Robert Legenstein
- Institute of Theoretical Computer Science, Graz University of TechnologyGrazAustria
| | - Wolfgang Maass
- Institute of Theoretical Computer Science, Graz University of TechnologyGrazAustria
| |
Collapse
|
38
|
Wunderlich TC, Pehle C. Event-based backpropagation can compute exact gradients for spiking neural networks. Sci Rep 2021; 11:12829. [PMID: 34145314 PMCID: PMC8213775 DOI: 10.1038/s41598-021-91786-z] [Citation(s) in RCA: 13] [Impact Index Per Article: 4.3] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/13/2020] [Accepted: 05/28/2021] [Indexed: 11/09/2022] Open
Abstract
Spiking neural networks combine analog computation with event-based communication using discrete spikes. While the impressive advances of deep learning are enabled by training non-spiking artificial neural networks using the backpropagation algorithm, applying this algorithm to spiking networks was previously hindered by the existence of discrete spike events and discontinuities. For the first time, this work derives the backpropagation algorithm for a continuous-time spiking neural network and a general loss function by applying the adjoint method together with the proper partial derivative jumps, allowing for backpropagation through discrete spike events without approximations. This algorithm, EventProp, backpropagates errors at spike times in order to compute the exact gradient in an event-based, temporally and spatially sparse fashion. We use gradients computed via EventProp to train networks on the Yin-Yang and MNIST datasets using either a spike time or voltage based loss function and report competitive performance. Our work supports the rigorous study of gradient-based learning algorithms in spiking neural networks and provides insights toward their implementation in novel brain-inspired hardware.
Collapse
Affiliation(s)
- Timo C Wunderlich
- Kirchhoff-Institute for Physics, Heidelberg University, 69120, Heidelberg, Germany.
- Berlin Institute of Health, Charité-Universitätsmedizin, 10117, Berlin, Germany.
| | - Christian Pehle
- Kirchhoff-Institute for Physics, Heidelberg University, 69120, Heidelberg, Germany.
| |
Collapse
|
39
|
Gardner B, Grüning A. Supervised Learning With First-to-Spike Decoding in Multilayer Spiking Neural Networks. Front Comput Neurosci 2021; 15:617862. [PMID: 33912021 PMCID: PMC8072060 DOI: 10.3389/fncom.2021.617862] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/15/2020] [Accepted: 03/08/2021] [Indexed: 11/18/2022] Open
Abstract
Experimental studies support the notion of spike-based neuronal information processing in the brain, with neural circuits exhibiting a wide range of temporally-based coding strategies to rapidly and efficiently represent sensory stimuli. Accordingly, it would be desirable to apply spike-based computation to tackling real-world challenges, and in particular transferring such theory to neuromorphic systems for low-power embedded applications. Motivated by this, we propose a new supervised learning method that can train multilayer spiking neural networks to solve classification problems based on a rapid, first-to-spike decoding strategy. The proposed learning rule supports multiple spikes fired by stochastic hidden neurons, and yet is stable by relying on first-spike responses generated by a deterministic output layer. In addition to this, we also explore several distinct, spike-based encoding strategies in order to form compact representations of presented input data. We demonstrate the classification performance of the learning rule as applied to several benchmark datasets, including MNIST. The learning rule is capable of generalizing from the data, and is successful even when used with constrained network architectures containing few input and hidden layer neurons. Furthermore, we highlight a novel encoding strategy, termed "scanline encoding," that can transform image data into compact spatiotemporal patterns for subsequent network processing. Designing constrained, but optimized, network structures and performing input dimensionality reduction has strong implications for neuromorphic applications.
Collapse
Affiliation(s)
- Brian Gardner
- Department of Computer Science, University of Surrey, Guildford, United Kingdom
| | - André Grüning
- Faculty of Electrical Engineering and Computer Science, University of Applied Sciences, Stralsund, Germany
| |
Collapse
|