Abstract:Various hypotheses of information representation in brain, referred to as neural codes, have been proposed to explain the information transmission between neurons. Neural coding plays an essential role in enabling the brain-inspired spiking neural networks (SNNs) to perform different tasks. To search for the best coding scheme, we performed an extensive comparative study on the impact and performance of four important neural coding schemes, namely, rate coding, time-to-first spike (TTFS) coding, phase coding, and burst coding. The comparative study was carried out using a biological 2-layer SNN trained with an unsupervised spike-timing-dependent plasticity (STDP) algorithm. Various aspects of network performance were considered, including classification accuracy, processing latency, synaptic operations (SOPs), hardware implementation, network compression efficacy, input and synaptic noise resilience, and synaptic fault tolerance. The classification tasks on Modified National Institute of Standards and Technology (MNIST) and Fashion-MNIST datasets were applied in our study. For hardware implementation, area and power consumption were estimated for these coding schemes, and the network compression efficacy was analyzed using pruning and quantization techniques. Different types of input noise and noise variations in the datasets were considered and applied. Furthermore, the robustness of each coding scheme to the non-ideality-induced synaptic noise and fault in analog neuromorphic systems was studied and compared. Our results show that TTFS coding is the best choice in achieving the highest computational performance with very low hardware implementation overhead. TTFS coding requires 4x/7.5x lower processing latency and 3.5x/6.5x fewer SOPs than rate coding during the training/inference process. Phase coding is the most resilient scheme to input noise. Burst coding offers the highest network compression efficacy and the best overall robustness to hardware non-idealities for both training and inference processes. The study presented in this paper reveals the design space created by the choice of each coding scheme, allowing designers to frame each scheme in terms of its strength and weakness given a designs’ constraints and considerations in neuromorphic systems.

Robust Transcoding Sensory Information with Neural Spikes

Robust Sensory Information Reconstruction and Classification With Augmented Spikes

Sparse Temporal Encoding of Visual Features for Robust Object Recognition by Spiking Neurons

Robust Decoding of Rich Dynamical Visual Scenes With Retinal Spikes

Hierarchical Spiking-Based Model for Efficient Image Classification with Enhanced Feature Extraction and Encoding.

A Multi-Spike Approach For Robust Sound Recognition

Toward Efficient Processing and Learning with Spikes: New Approaches for Multispike Learning

A Spiking Neural Network System for Robust Sequence Recognition

Motorsrnn: A Spiking Recurrent Neural Network Inspired by Brain Topology for the Effective and Efficient Decoding of Cortical Spike Trains

Binless Kernel Machine: Modeling Spike Train Transformation for Cognitive Neural Prostheses

Deep CovDenseSNN: A Hierarchical Event-Driven Dynamic Framework with Spiking Neurons in Noisy Environment

CSNN: an Augmented Spiking Based Framework with Perceptron-Inception

Reconstruction of natural visual scenes from neural spikes with deep neural networks

An Artificial Visual Neuron with Multiplexed Rate and Time-to-first-spike Coding

Neural encoding with unsupervised spiking convolutional neural network

A Temporal Encoding Method Based on Expansion Representation

Neuromorphic Auditory Perception by Neural Spiketrum

Selective and Compressive Sensing for Energy-Efficient Implantable Neural Decoding.

Neural Coding in Spiking Neural Networks: A Comparative Study for Robust Neuromorphic Systems

A Brain-Inspired Spiking Neural Network Model with Temporal Encoding and Learning

Neural stochastic codes, encoding and decoding