Simple and complex spiking neurons: perspectives and analysis in a simple STDP scenario

Abstract

Spiking neural networks (SNNs) are largely inspired by biology and neuroscience, and leverage ideas and theories to create fast and efficient learning systems. Spiking neuron models are adopted as core processing units in neuromorphic systems because they enable event-based processing. The integrate-and-fire (I\&F) models are often adopted as considered more suitable, with the simple Leaky I\&F (LIF) being the most used. The reason for adopting such models is their efficiency or biological plausibility. Nevertheless, rigorous justification for the adoption of LIF over other neuron models for use in artificial learning systems has not yet been studied. This work considers a variety of neuron models in the literature and then selects computational neuron models that are single-variable, efficient, and display different types of complexities. From this selection, we make a comparative study of three simple I\&F neuron models, namely the LIF, the Quadratic I\&F (QIF) and the Exponential I\&F (EIF), to understand whether the use of more complex models increases the performance of the system and whether the choice of a neuron model can be directed by the task to be completed. Neuron models are tested within an SNN trained with Spike-Timing Dependent Plasticity (STDP) on a classification task on the N-MNIST and DVS Gestures datasets. Experimental results reveal that more complex neurons manifest the same ability as simpler ones to achieve high levels of accuracy on a simple dataset (N-MNIST), albeit requiring comparably more hyper-parameter tuning. However, when the data possess richer spatio-temporal features, the QIF and EIF neuron models steadily achieve better results. This suggests that accurately selecting the model based on the richness of the feature spectrum of the data could improve the performance of the whole system. Finally, the code implementing the spiking neurons in the SpykeTorch framework is made publicly available.

Molecular Toxicity Virtual Screening Applying a Quantized Computational SNN-Based Framework

“…Each SNN had several hyperparameters (i.e., HPs) to be tuned: in the previous literature [ 72 , 73 , 74 ], this part is usually reported as time-consuming and challenging to perform due to the non-linear relationship between LIF output and HPs. In the current investigation, nested cross-validation (i.e., CV) has been employed to separate the phase of HP optimization and model evaluation.…”

From fundamentals to frontiers: a review of memristor mechanisms, modeling and emerging applications

“…101,102 The preferred kind of neuron may depend on the requirements of data and type of computation. 103 There is great interest in building biology-like neurons by using different materials platforms, such as nano fluidics 104 and organic and inorganic semiconductors. 105,106…”

Computing of neuromorphic materials: an emerging approach for bioengineering solutions

“…These possibilities are being investigated by ongoing research, although it might be some time before a completely co-designed neuromorphic computing stack is used in everyday life. [340][341][342] This is because these tasks require complex algorithms and architectures to be implemented, and these algorithms are difficult to design and optimize in neuromorphic computers. Additionally, it is hard to make sure that the entire stack works together seamlessly.…”