Abstract:Spiking neural network (SNN) is a brain-inspired model with more spatio-temporal information processing capacity and computational energy efficiency. However, with the increasing depth of SNNs, the memory problem caused by the weights of SNNs has gradually attracted attention. In this study, we propose an ultra-low latency adaptive local binary spiking neural network (ALBSNN) with accuracy loss estimators, which dynamically selects the network layers to be binarized to ensure a balance between quantization degree and classification accuracy by evaluating the error caused by the binarized weights during the network learning process. At the same time, to accelerate the training speed of the network, the global average pooling (GAP) layer is introduced to replace the fully connected layers by combining convolution and pooling. Finally, to further reduce the error caused by the binary weight, we propose binary weight optimization (BWO), which updates the overall weight by directly adjusting the binary weight. This method further reduces the loss of the network that reaches the training bottleneck. The combination of the above methods balances the network's quantization and recognition ability, enabling the network to maintain the recognition capability equivalent to the full precision network and reduce the storage space by more than 20%. So, SNNs can use a small number of time steps to obtain better recognition accuracy. In the extreme case of using only a one-time step, we still can achieve 93.39, 92.12, and 69.55% testing accuracy on three traditional static datasets, Fashion- MNIST, CIFAR-10, and CIFAR-100, respectively. At the same time, we evaluate our method on neuromorphic N-MNIST, CIFAR10-DVS, and IBM DVS128 Gesture datasets and achieve advanced accuracy in SNN with binary weights. Our network has greater advantages in terms of storage resources and training time.

NUTS-BSNN: A Non-uniform Time-step Binarized Spiking Neural Networks with Energy-Efficient In-memory Computing Macro

BitSNNs: Revisiting Energy-efficient Spiking Neural Networks

Ultra-low Latency Adaptive Local Binary Spiking Neural Network with Accuracy Loss Estimator

ALBSNN: ultra-low latency adaptive local binary spiking neural network with accuracy loss estimator

sBSNN: Stochastic-Bits Enabled Binary Spiking Neural Network with On-Chip Learning for Energy Efficient Neuromorphic Computing at the Edge

NBSSN: A Neuromorphic Binary Single-Spike Neural Network for Efficient Edge Intelligence.

Direct Training of Hardware-Friendly Weight Binarized Spiking Neural Network with Surrogate Gradient Learning Towards Spatio-Temporal Event-Based Dynamic Data Recognition

Deep Spiking Convolutional Neural Networks for Programmable Neuro-synaptic System

Improving energy efficiency and classification accuracy of neuromorphic chips by learning binary synaptic crossbars

A Scatter-and-Gather Spiking Convolutional Neural Network on a Reconfigurable Neuromorphic Hardware

Memristor-Based Binarized Spiking Neural Networks: Challenges and applications

Enabling Deep Spiking Neural Networks with Hybrid Conversion and Spike Timing Dependent Backpropagation

Batch Normalization-Free Weight-Binarized SNN Based on Hardware-Saving IF Neuron.

Deep Spiking Neural Network with Ternary Spikes

Training Energy-Efficient Deep Spiking Neural Networks with Single-Spike Hybrid Input Encoding

Efficient Discrete Temporal Coding Spike-Driven In-Memory Computing Macro for Deep Neural Network Based on Nonvolatile Memory.

You Only Spike Once: Improving Energy-Efficient Neuromorphic Inference to ANN-Level Accuracy

STBNN: Hardware-friendly Spatio-Temporal Binary Neural Network with High Pattern Recognition Accuracy

MT-Spike: A Multilayer Time-based Spiking Neuromorphic Architecture with Temporal Error Backpropagation

Ultra-low Latency Spiking Neural Networks with Spatio-Temporal Compression and Synaptic Convolutional Block

Improving classification accuracy of feedforward neural networks for spiking neuromorphic chips