SLAYER: Spike Layer Error Reassignment in Time

Abstract

Configuring deep Spiking Neural Networks (SNNs) is an exciting research avenue for low power spike event based computation. However, the spike generation function is non-differentiable and therefore not directly compatible with the standard error backpropagation algorithm. In this paper, we introduce a new general backpropagation mechanism for learning synaptic weights and axonal delays which overcomes the problem of non-differentiability of the spike function and uses a temporal credit assignment policy for backpropagating error to preceding layers. We describe and release a GPU accelerated software implementation of our method which allows training both fully connected and convolutional neural network (CNN) architectures. Using our software, we compare our method against existing SNN based learning approaches and standard ANN to SNN conversion techniques and show that our method achieves state of the art performance for an SNN on the MNIST, NMNIST, DVS Gesture, and TIDIGITS datasets.

arXiv comment10 pages, 2 figures

Cited by in corpus (83)

• Enabling Spike-based Backpropagation for Training Deep Neural Network Architectures
• Synaptic Plasticity Dynamics for Deep Continuous Local Learning (DECOLLE)
• The Heidelberg spiking datasets for the systematic evaluation of spiking neural networks
• Deep Residual Learning in Spiking Neural Networks
• Hardware and Software Optimizations for Accelerating Deep Neural Networks: Survey of Current Trends, Challenges, and the Road Ahead
• Unsupervised Learning of a Hierarchical Spiking Neural Network for Optical Flow Estimation: From Events to Global Motion Perception
• Surrogate Gradient Learning in Spiking Neural Networks
• Learning to Detect Objects with a 1 Megapixel Event Camera
• SpykeTorch: Efficient Simulation of Convolutional Spiking Neural Networks with at most one Spike per Neuron
• Temporal Spike Sequence Learning via Backpropagation for Deep Spiking Neural Networks
• Bottom-up and top-down approaches for the design of neuromorphic processing systems: Tradeoffs and synergies between natural and artificial intelligence
• Optimal Conversion of Conventional Artificial Neural Networks to Spiking Neural Networks
• DIET-SNN: Direct Input Encoding With Leakage and Threshold Optimization in Deep Spiking Neural Networks
• BS4NN: Binarized Spiking Neural Networks with Temporal Coding and Learning
• Spiking Deep Residual Network
• Is Neuromorphic MNIST neuromorphic? Analyzing the discriminative power of neuromorphic datasets in the time domain
• Event-based Gesture Recognition with Dynamic Background Suppression using Smartphone Computational Capabilities
• Constructing Accurate and Efficient Deep Spiking Neural Networks with Double-threshold and Augmented Schemes
• STSC-SNN: Spatio-Temporal Synaptic Connection with Temporal Convolution and Attention for Spiking Neural Networks
• Enabling Deep Spiking Neural Networks with Hybrid Conversion and Spike Timing Dependent Backpropagation
• The Backpropagation Algorithm Implemented on Spiking Neuromorphic Hardware
• RMP-SNN: Residual Membrane Potential Neuron for Enabling Deeper High-Accuracy and Low-Latency Spiking Neural Network
• Quantum Superposition Inspired Spiking Neural Network
• Spiking neural networks trained with backpropagation for low power neuromorphic implementation of voice activity detection
• Keys to Accurate Feature Extraction Using Residual Spiking Neural Networks
• Training Deep Spiking Neural Networks
• Spiking Neural Networks for Detecting Satellite-Based Internet-of-Things Signals
• An Energy-Efficient Spiking Neural Network for Finger Velocity Decoding for Implantable Brain-Machine Interface
• DeepFire2: A Convolutional Spiking Neural Network Accelerator on FPGAs
• Spiking neural networks trained via proxy
• Integration of Neuromorphic AI in Event-Driven Distributed Digitized Systems: Concepts and Research Directions
• Learning Delays in Spiking Neural Networks using Dilated Convolutions with Learnable Spacings
• A Tandem Learning Rule for Effective Training and Rapid Inference of Deep Spiking Neural Networks
• Empirical study on the efficiency of Spiking Neural Networks with axonal delays, and algorithm-hardware benchmarking
• Comparing SNNs and RNNs on Neuromorphic Vision Datasets: Similarities and Differences
• Event Transformer+. A multi-purpose solution for efficient event data processing
• Progressive Tandem Learning for Pattern Recognition with Deep Spiking Neural Networks
• Reinforcement co-Learning of Deep and Spiking Neural Networks for Energy-Efficient Mapless Navigation with Neuromorphic Hardware
• SPIDE: A Purely Spike-based Method for Training Feedback Spiking Neural Networks
• Parallel Hyperparameter Optimization Of Spiking Neural Network
• Automotive Object Detection via Learning Sparse Events by Spiking Neurons
• SpikePropamine: Differentiable Plasticity in Spiking Neural Networks
• Effective and Efficient Computation with Multiple-timescale Spiking Recurrent Neural Networks
• Finite Meta-Dynamic Neurons in Spiking Neural Networks for Spatio-temporal Learning
• Special Session: Neuromorphic hardware design and reliability from traditional CMOS to emerging technologies
• Exploiting Neuron and Synapse Filter Dynamics in Spatial Temporal Learning of Deep Spiking Neural Network
• Encoding Event-Based Data With a Hybrid SNN Guided Variational Auto-encoder in Neuromorphic Hardware
• A Spike Learning System for Event-driven Object Recognition
• Unifying Activation- and Timing-based Learning Rules for Spiking Neural Networks
• Efficient Deep Spiking Multi-Layer Perceptrons with Multiplication-Free Inference
• DCT-SNN: Using DCT to Distribute Spatial Information over Time for Learning Low-Latency Spiking Neural Networks
• Training Feedback Spiking Neural Networks by Implicit Differentiation on the Equilibrium State
• A Low Latency Adaptive Coding Spiking Framework for Deep Reinforcement Learning
• Spike-based computation using classical recurrent neural networks
• Event-Driven Visual-Tactile Sensing and Learning for Robots
• Energy efficiency analysis of Spiking Neural Networks for space applications
• MAP-SNN: Mapping Spike Activities with Multiplicity, Adaptability, and Plasticity into Bio-Plausible Spiking Neural Networks
• An efficient and perceptually motivated auditory neural encoding and decoding algorithm for spiking neural networks
• You Only Spike Once: Improving Energy-Efficient Neuromorphic Inference to ANN-Level Accuracy
• Optimising Event-Driven Spiking Neural Network with Regularisation and Cutoff
• Sparse Spiking Gradient Descent
• One Timestep is All You Need: Training Spiking Neural Networks with Ultra Low Latency
• One-Shot Federated Learning with Neuromorphic Processors
• Ensembles of Spiking Neural Networks
• ST-MNIST -- The Spiking Tactile MNIST Neuromorphic Dataset
• TactileSGNet: A Spiking Graph Neural Network for Event-based Tactile Object Recognition
• Combining Spiking Neural Network and Artificial Neural Network for Enhanced Image Classification
• Neuromorphic Processing and Sensing: Evolutionary Progression of AI to Spiking
• Learning to Time-Decode in Spiking Neural Networks Through the Information Bottleneck
• Attacking the Spike: On the Transferability and Security of Spiking Neural Networks to Adversarial Examples
• Multivariate Time Series Classification Using Spiking Neural Networks
• A temporally and spatially local spike-based backpropagation algorithm to enable training in hardware
• Minibatch Processing in Spiking Neural Networks
• Harnessing Slow Dynamics in Neuromorphic Computation
• SPENSER: Towards a NeuroEvolutionary Approach for Convolutional Spiking Neural Networks
• Bifurcation Spiking Neural Network
• BioLeaF: A Bio-plausible Learning Framework for Training of Spiking Neural Networks
• Multi-domain Collaborative Feature Representation for Robust Visual Object Tracking
• 2022 Roadmap on Neuromorphic Computing and Engineering
• Data-Driven Neuromorphic DRAM-based CNN and RNN Accelerators
• WaveSense: Efficient Temporal Convolutions with Spiking Neural Networks for Keyword Spotting
• On-Chip Error-triggered Learning of Multi-layer Memristive Spiking Neural Networks
• Temporal Surrogate Back-propagation for Spiking Neural Networks