Abstract:<p>Machine learning is yielding unprecedented interest in research and industry, due to recent success in many applied contexts such as image classification and object recognition. However, the deployment of these systems requires huge computing capabilities, thus making them unsuitable for embedded systems. To deal with this limitation, many researchers are investigating brain-inspired computing, which would be a perfect alternative to the conventional Von Neumann architecture based computers (CPU/GPU) that meet the requirements for computing performance, but not for energy-efficiency. Therefore, neuromorphic hardware circuits that are adaptable for both parallel and distributed computations need to be designed. In this paper, we focus on Spiking Neural Networks (SNNs) with a comprehensive study of information coding methods and hardware exploration. In this context, we propose a framework for neuromorphic hardware design space exploration, which allows to define a suitable architecture based on application-specific constraints and starting from a wide variety of possible architectural choices. For this framework, we have developed a behavioral level simulator for neuromorphic hardware architectural exploration named NAXT. Moreover, we propose modified versions of the standard Rate Coding technique to make trade-offs with the Time Coding paradigm, which is characterized by the low number of spikes propagating in the network. Thus, we are able to reduce the number of spikes while keeping the same neuron's model, which results in an SNN with fewer events to process. By doing so, we seek to reduce the amount of power consumed by the hardware. Furthermore, we present three neuromorphic hardware architectures in order to quantitatively study the implementation of SNNs. One of these architectures integrates a novel hybrid structure: a highly-parallel computation core for most solicited layers, and time-multiplexed computation units for deeper layers. These architectures are derived from a novel funnel-like Design Space Exploration framework for neuromorphic hardware.</p>

Spiking recurrent neural networks for neuromorphic computing in nonlinear structural mechanics

Brain-inspired spiking neural networks in Engineering Mechanics: a new physics-based self-learning framework for sustainable Finite Element analysis

Spiking neural networks for nonlinear regression

Exploring Neuromorphic Computing Based on Spiking Neural Networks: Algorithms to Hardware

Physics-Based Self-Learning Spiking Neural Network enhanced time-integration scheme for computing viscoplastic structural finite element response

Recent Advances in Scalable Energy-Efficient and Trustworthy Spiking Neural networks: from Algorithms to Technology

Design Space Exploration of Hardware Spiking Neurons for Embedded Artificial Intelligence

Spiking Neurons with Neural Dynamics Implemented Using Stochastic Memristors

Spiking Neural Networks for Structural Health Monitoring

Advancements in Algorithms and Neuromorphic Hardware for Spiking Neural Networks

Neuromorphic Processing and Sensing: Evolutionary Progression of AI to Spiking

Conversion of Artificial Recurrent Neural Networks to Spiking Neural Networks for Low-power Neuromorphic Hardware

Compiling Spiking Neural Networks to Neuromorphic Hardware

Low-Power Neuromorphic Hardware for Signal Processing Applications

Efficient Implementation of Spiking Neural Networks for Inference Using Ex-Situ Training

Experimental Demonstration of Supervised Learning in Spiking Neural Networks with Phase-Change Memory Synapses

In-Hardware Learning of Multilayer Spiking Neural Networks on a Neuromorphic Processor

A Long Short-Term Memory for AI Applications in Spike-based Neuromorphic Hardware

PT-Spike: A Precise-Time-Dependent Single Spike Neuromorphic Architecture with Efficient Supervised Learning

To Spike or Not To Spike: A Digital Hardware Perspective on Deep Learning Acceleration

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