Abstract:The problem of training spiking neural networks (SNNs) is a necessary precondition to understanding computations within the brain, a field still in its infancy. Previous work has shown that supervised learning in multi-layer SNNs enables bio-inspired networks to recognize patterns of stimuli through hierarchical feature acquisition. Although gradient descent has shown impressive performance in multi-layer (and deep) SNNs, it is generally not considered biologically plausible and is also computationally expensi… Show more
“…In [118], a supervised learning method was proposed (BP-STDP) where the backpropagation update rules were converted to temporally local STDP rules for multilayer SNNs. This model achieved accuracies comparable to equal-sized conventional and spiking networks for the MNIST benchmark (see Section III-A).…”
1 In recent years, deep learning has revolutionized the field of machine learning, for computer vision in particular. In this approach, a deep (multilayer) artificial neural network (ANN) is trained in a supervised manner using backpropagation. Vast amounts of labeled training examples are required, but the resulting classification accuracy is truly impressive, sometimes outperforming humans. Neurons in an ANN are characterized by a single, static, continuous-valued activation. Yet biological neurons use discrete spikes to compute and transmit information, and the spike times, in addition to the spike rates, matter. Spiking neural networks (SNNs) are thus more biologically realistic than ANNs, and arguably the only viable option if one wants to understand how the brain computes. SNNs are also more hardware friendly and energy-efficient than ANNs, and are thus appealing for technology, especially for portable devices. However, training deep SNNs remains a challenge. Spiking neurons' transfer function is usually non-differentiable, which prevents using backpropagation.Here we review recent supervised and unsupervised methods to train deep SNNs, and compare them in terms of accuracy, but also computational cost and hardware friendliness. The emerging picture is that SNNs still lag behind ANNs in terms of accuracy, but the gap is decreasing, and can even vanish on some tasks, while SNNs typically require many fewer operations.
“…In [118], a supervised learning method was proposed (BP-STDP) where the backpropagation update rules were converted to temporally local STDP rules for multilayer SNNs. This model achieved accuracies comparable to equal-sized conventional and spiking networks for the MNIST benchmark (see Section III-A).…”
1 In recent years, deep learning has revolutionized the field of machine learning, for computer vision in particular. In this approach, a deep (multilayer) artificial neural network (ANN) is trained in a supervised manner using backpropagation. Vast amounts of labeled training examples are required, but the resulting classification accuracy is truly impressive, sometimes outperforming humans. Neurons in an ANN are characterized by a single, static, continuous-valued activation. Yet biological neurons use discrete spikes to compute and transmit information, and the spike times, in addition to the spike rates, matter. Spiking neural networks (SNNs) are thus more biologically realistic than ANNs, and arguably the only viable option if one wants to understand how the brain computes. SNNs are also more hardware friendly and energy-efficient than ANNs, and are thus appealing for technology, especially for portable devices. However, training deep SNNs remains a challenge. Spiking neurons' transfer function is usually non-differentiable, which prevents using backpropagation.Here we review recent supervised and unsupervised methods to train deep SNNs, and compare them in terms of accuracy, but also computational cost and hardware friendliness. The emerging picture is that SNNs still lag behind ANNs in terms of accuracy, but the gap is decreasing, and can even vanish on some tasks, while SNNs typically require many fewer operations.
“…In Mostafa (2017) [10], the use of 800 IF neurons with alpha functions complicates the neural processing and the learning procedure of the network. In Tavanaei et al (2018) [15], the network's computational cost is quite large due to the use of rate coding and 1000 hidden neurons. In Comsa et al (2019) [11], the use of complicated SRM neuron model with exponential synaptic current makes it difficult for event-based implementation.…”
We propose a new supervised learning rule for multilayer spiking neural networks (SNN) that use a form of temporal coding known as rank-ordercoding. With this coding scheme, all neurons fire exactly one spike per stimulus, but the firing order carries information. In particular, in the readout layer the first neuron to fire determines the class of the stimulus. We derive a new learning rule for this sort of network, termed S4NN, akin to traditional error backpropagation, yet based on latencies. We show how approximate error gradients can be computed backward in a feedforward network with an arbitrary number of layers. This approach reaches state-of-the-art performance with SNNs: test accuracy of 97.4% for the MNIST dataset, and of 99.2% for the Caltech Face/Motorbike dataset. Yet the neuron model we use, non-leaky integrate-and-fire, are simpler and more hardware friendly than the one used in all previous similar proposals.
“…Current SNN models for pattern recognition can be generally categorized into three classes: that is, indirect training [12,13,14,15,16,17,18,19], direct SL training with BP [11,26,20,21,22,23,53], and plasticity-based unsupervised training with supervised modules [54,24,25]. for optimal initial weights and then used current-based BP to re-train all-layer weights in a supervised way [53]; however, this also resulted in the model being bio-implausible due to the use of the BP algorithm.…”
Section: Comparison With Other Snn Modelsmentioning
confidence: 99%
“…For the indirect SL method, ANNs are first trained and then mapped to equivalent SNNs by different conversion algorithms that transform real-valued computing into spike-based computing [12,13,14,15,16,17,18,19]; however, this method does not incorporate SNN learning and therefore provides no heuristic information on how to train a SNN. The direct SL method is based on the BP algorithm [11,20,21,22,23], e.g., using membrane potentials as continuous variables for calculating errors in BP [20,23] or using continuous activity function to approximate neuronal spike activity and obtain differentiable activ-ity for the BP algorithm [11,22]. However, such research must still perform numerous real-valued computations and non-local communications during the training process; thus, BP-based methods are as potentially energy inefficient as ANNs and also lack bio-plausibility.…”
Spiking neural networks (SNNs) possess energy-efficient potential due to eventbased computation. However, supervised training of SNNs remains a challenge as spike activities are non-differentiable. Previous SNNs training methods can be generally categorized into two basic classes, i.e., backpropagation-like training methods and plasticity-based learning methods. The former methods are dependent on energy-inefficient real-valued computation and non-local transmission, as also required in artificial neural networks (ANNs), whereas the latter are either considered to be biologically implausible or exhibit poor performance.Hence, biologically plausible (bio-plausible) high-performance supervised learning (SL) methods for SNNs remain deficient. In this paper, we proposed a novel bio-plausible SNN model for SL based on the symmetric spike-timing dependent plasticity (sym-STDP) rule found in neuroscience. By combining the sym-STDP rule with bio-plausible synaptic scaling and intrinsic plasticity of the dynamic threshold, our SNN model implemented SL well and achieved good performance in the benchmark recognition task (MNIST dataset). To reveal the underlying mechanism of our SL model, we visualized both layer-based * These two authors contributed equally to this work.activities and synaptic weights using the t-distributed stochastic neighbor embedding (t-SNE) method after training and found that they were well clustered, thereby demonstrating excellent classification ability. Furthermore, to verify the robustness of our model, we trained it on another more realistic dataset (Fashion-MNIST), which also showed good performance. As the learning rules were bio-plausible and based purely on local spike events, our model could be easily applied to neuromorphic hardware for online training and may be helpful for understanding SL information processing at the synaptic level in biological neural systems.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.