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.

Adaptive Linear Unit for Accurate Binary Neural Networks

AdaBin: Improving Binary Neural Networks with Adaptive Binary Sets

Improving Fault Tolerance for Reliable DNN Using Boundary-Aware Activation

Shallow Binary Features Enhance the Robustness of Deep Convolutional Neural Networks

LDF-BNN: A Real-Time and High-Accuracy Binary Neural Network Accelerator Based on the Improved BNext

Bi-Real Net V2: Rethinking Non-linearity for 1-Bit CNNs and Going Beyond

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

SA-BNN: State-Aware Binary Neural Network.

Universal Binary Neural Networks Design by Improved Differentiable Neural Architecture Search

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

Rad-Bnn: Regulating Activation Distribution for Accurate Binary Neural Network

Learning Adaptive Gradients for Binary Neural Networks

FTBNN: Rethinking Non-linearity for 1-bit CNNs and Going Beyond

Recurrent Bilinear Optimization for Binary Neural Networks

RAD-BNN

A&B BNN: Add&Bit-Operation-Only Hardware-Friendly Binary Neural Network

Network Binarization via Contrastive Learning

Towards Accurate Binary Neural Networks via Modeling Contextual Dependencies.

IE-Net: Information-Enhanced Binary Neural Networks for Accurate Classification

Resilient Binary Neural Network.

Training Multi-Layer Binary Neural Networks With Local Binary Error Signals