Exploiting High Performance Spiking Neural Networks with Efficient Spiking Patterns

Spiking Neural Networks (SNNs) use discrete spike sequences to transmit information, which significantly mimics the information transmission of the brain. Although this binarized form of representation dramatically enhances the energy efficiency and robustness of SNNs, it also leaves a large gap between the performance of SNNs and Artificial Neural Networks based on real values. There are many different spike patterns in the brain, and the dynamic synergy of these spike patterns greatly enriches the representation capability. Inspired by spike patterns in biological neurons, this paper introduces the dynamic Burst pattern and designs the Leaky Integrate and Fire or Burst (LIFB) neuron that can make a trade-off between short-time performance and dynamic temporal performance from the perspective of network information capacity. LIFB neuron exhibits three modes, resting, Regular spike, and Burst spike. The burst density of the neuron can be adaptively adjusted, which significantly enriches the characterization capability. We also propose a decoupling method that can losslessly decouple LIFB neurons into equivalent LIF neurons, which demonstrates that LIFB neurons can be efficiently implemented on neuromorphic hardware. We conducted experiments on the static datasets CIFAR10, CIFAR100, and ImageNet, which showed that we greatly improved the performance of the SNNs while significantly reducing the network latency. We also conducted experiments on neuromorphic datasets DVS-CIFAR10 and NCALTECH101 and showed that we achieved state-of-the-art with a small network structure.

[1]  Lei Deng,et al.  Attention Spiking Neural Networks , 2022, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[2]  Wei Fang,et al.  Optimal ANN-SNN Conversion for High-accuracy and Ultra-low-latency Spiking Neural Networks , 2023, ICLR.

[3]  Yiting Dong,et al.  BrainCog: A Spiking Neural Network based Brain-inspired Cognitive Intelligence Engine for Brain-inspired AI and Brain Simulation , 2022, SSRN Electronic Journal.

[4]  Yule Duan,et al.  TCJA-SNN: Temporal-Channel Joint Attention for Spiking Neural Networks , 2022, IEEE transactions on neural networks and learning systems.

[5]  Liwen Zhang,et al.  RecDis-SNN: Rectifying Membrane Potential Distribution for Directly Training Spiking Neural Networks , 2022, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[6]  Shenmin Zhang,et al.  Gradual Surrogate Gradient Learning in Deep Spiking Neural Networks , 2022, ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[7]  Yi Zeng,et al.  Efficient and Accurate Conversion of Spiking Neural Network with Burst Spikes , 2022, IJCAI.

[8]  Shanghang Zhang,et al.  Temporal Efficient Training of Spiking Neural Network via Gradient Re-weighting , 2022, ICLR.

[9]  Sungroh Yoon,et al.  AutoSNN: Towards Energy-Efficient Spiking Neural Networks , 2022, ICML.

[10]  P. Panda,et al.  Neural Architecture Search for Spiking Neural Networks , 2022, ECCV.

[11]  Jihang Wang,et al.  Spiking CapsNet: A Spiking Neural Network With A Biologically Plausible Routing Rule Between Capsules , 2021, Inf. Sci..

[12]  Dongcheng Zhao,et al.  Backpropagation with biologically plausible spatiotemporal adjustment for training deep spiking neural networks , 2021, Patterns.

[13]  Yi Zeng,et al.  BackEISNN: A Deep Spiking Neural Network with Adaptive Self-Feedback and Balanced Excitatory-Inhibitory Neurons , 2021, Neural Networks.

[14]  Yi Zeng,et al.  BSNN: Towards faster and better conversion of artificial neural networks to spiking neural networks with bistable neurons , 2021, Frontiers in Neuroscience.

[15]  Guoqi Li,et al.  LIAF-Net: Leaky Integrate and Analog Fire Network for Lightweight and Efficient Spatiotemporal Information Processing , 2020, IEEE Transactions on Neural Networks and Learning Systems.

[16]  Kay Chen Tan,et al.  Constructing Accurate and Efficient Deep Spiking Neural Networks With Double-Threshold and Augmented Schemes , 2020, IEEE Transactions on Neural Networks and Learning Systems.

[17]  K. Roy,et al.  DIET-SNN: A Low-Latency Spiking Neural Network With Direct Input Encoding and Leakage and Threshold Optimization , 2021, IEEE Transactions on Neural Networks and Learning Systems.

[18]  Guoqi Li,et al.  Temporal-wise Attention Spiking Neural Networks for Event Streams Classification , 2021, 2021 IEEE/CVF International Conference on Computer Vision (ICCV).

[19]  Lei Deng,et al.  Exploiting Spiking Dynamics with Spatial-temporal Feature Normalization in Graph Learning , 2021, IJCAI.

[20]  Malu Zhang,et al.  Deep Spiking Neural Network with Neural Oscillation and Spike-Phase Information , 2021, AAAI.

[21]  Sander M. Bohte,et al.  Accurate and efficient time-domain classification with adaptive spiking recurrent neural networks , 2021, Nature Machine Intelligence.

[22]  Yonghong Tian,et al.  Deep Residual Learning in Spiking Neural Networks , 2021, NeurIPS.

[23]  Yujie Wu,et al.  Going Deeper With Directly-Trained Larger Spiking Neural Networks , 2020, AAAI.

[24]  Peng Li,et al.  Skip-Connected Self-Recurrent Spiking Neural Networks With Joint Intrinsic Parameter and Synaptic Weight Training , 2020, Neural Computation.

[25]  P. Panda,et al.  Revisiting Batch Normalization for Training Low-Latency Deep Spiking Neural Networks From Scratch , 2020, Frontiers in Neuroscience.

[26]  Yonghong Tian,et al.  Incorporating Learnable Membrane Time Constant to Enhance Learning of Spiking Neural Networks , 2020, 2021 IEEE/CVF International Conference on Computer Vision (ICCV).

[27]  Idan Segev,et al.  Single cortical neurons as deep artificial neural networks , 2019, Neuron.

[28]  Shanghang Zhang,et al.  Differentiable Spike: Rethinking Gradient-Descent for Training Spiking Neural Networks , 2021, NeurIPS.

[29]  O. Papaspiliopoulos High-Dimensional Probability: An Introduction with Applications in Data Science , 2020 .

[30]  Kaushik Roy,et al.  DIET-SNN: Direct Input Encoding With Leakage and Threshold Optimization in Deep Spiking Neural Networks , 2020, ArXiv.

[31]  Bo Xu,et al.  LISNN: Improving Spiking Neural Networks with Lateral Interactions for Robust Object Recognition , 2020, IJCAI.

[32]  Elisabetta Chicca,et al.  Efficient Processing of Spatio-Temporal Data Streams With Spiking Neural Networks , 2020, Frontiers in Neuroscience.

[33]  Kaushik Roy,et al.  Enabling Deep Spiking Neural Networks with Hybrid Conversion and Spike Timing Dependent Backpropagation , 2020, ICLR.

[34]  Wenrui Zhang,et al.  Temporal Spike Sequence Learning via Backpropagation for Deep Spiking Neural Networks , 2020, NeurIPS.

[35]  Antoine Dupret,et al.  SpikeGrad: An ANN-equivalent Computation Model for Implementing Backpropagation with Spikes , 2019, ICLR.

[36]  Hong Yang,et al.  DART: Distribution Aware Retinal Transform for Event-Based Cameras , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[37]  Lei Deng,et al.  Direct Training for Spiking Neural Networks: Faster, Larger, Better , 2018, AAAI.

[38]  Sungroh Yoon,et al.  Fast and Efficient Information Transmission with Burst Spikes in Deep Spiking Neural Networks , 2018, 2019 56th ACM/IEEE Design Automation Conference (DAC).

[39]  Hao Li,et al.  Visualizing the Loss Landscape of Neural Nets , 2017, NeurIPS.

[40]  Lei Deng,et al.  Spatio-Temporal Backpropagation for Training High-Performance Spiking Neural Networks , 2017, Front. Neurosci..

[41]  Luping Shi,et al.  CIFAR10-DVS: An Event-Stream Dataset for Object Classification , 2017, Front. Neurosci..

[42]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[43]  Gregory Cohen,et al.  Converting Static Image Datasets to Spiking Neuromorphic Datasets Using Saccades , 2015, Front. Neurosci..

[44]  Tianqi Chen,et al.  Empirical Evaluation of Rectified Activations in Convolutional Network , 2015, ArXiv.

[45]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[46]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[47]  Alex Krizhevsky,et al.  Learning Multiple Layers of Features from Tiny Images , 2009 .

[48]  Eugene M. Izhikevich,et al.  Which model to use for cortical spiking neurons? , 2004, IEEE Transactions on Neural Networks.

[49]  Eugene M. Izhikevich,et al.  Simple model of spiking neurons , 2003, IEEE Trans. Neural Networks.

[50]  Frank C. Hoppensteadt,et al.  Bursts as a unit of neural information: selective communication via resonance , 2003, Trends in Neurosciences.

[51]  Adam Kepecs,et al.  Information encoding and computation with spikes and bursts , 2003, Network.

[52]  Peter Dayan,et al.  Theoretical Neuroscience: Computational and Mathematical Modeling of Neural Systems , 2001 .

[53]  S. Sherman,et al.  Fourier analysis of sinusoidally driven thalamocortical relay neurons and a minimal integrate-and-fire-or-burst model. , 2000, Journal of neurophysiology.

[54]  J. Lisman Bursts as a unit of neural information: making unreliable synapses reliable , 1997, Trends in Neurosciences.

[55]  A. Hodgkin,et al.  A quantitative description of membrane current and its application to conduction and excitation in nerve , 1990, Bulletin of mathematical biology.

[56]  R. Winder Partitions of N-Space by Hyperplanes , 1966 .