Design-Space Exploration of SNN Models Using Application-Specific Multi-Core Architectures

EasyChair Preprint 11905

Abstract

The most cutting-edge SNN simulators, for example, Brian2, NEST, and CARLsim were designed primarily for studying brain functions and neuronal dynamics, but due to the use of certain simulators, these tools are not user-friendly, additional functions must be specified in low-level programming languages like C++ and integrated into the simulator code. Such simulators allow the users to acquire simulations precisely in a relatively short period as compared to ANNs. Nevertheless, there are many challenges and computational issues related to SNN. For example, the model must initialize with the accurate biological representations of the neurons. However, the present simulators require a lot of time and extensive amounts of code for constructing the neural network designs as well as for evaluating and visualizing their behavior, hence an interactive simulator with little to no code is needed given all these limitations. Furthermore, a fast, run-time interaction, visualization, and analysis-based simulator not only helps to accelerate the simulation process but will also speed up the designing, prototyping, and parameter tuning. Additionally, the area as a whole also benefits from the study of other algorithms, such as training advances in the whole field. With this motivation and the difficulties that currently exist in comprehending and utilizing the promising features of SNNs, we proposed a novel run-time multi-core architecture-based simulator called "RAVSim" (Runtime Analysis and Visualization Simulator), a cutting-edge SNN simulator, developed using LabVIEW and it is publicly available on their website as an official module. RAVSim is a runtime virtual simulation environment tool that enables the user to interact with the model, observe its behavior of output concentration, and modify the set of parametric values at any time while the simulation is in execution.

Keyphrases: Computing methodologies, Runtime Simulator, Spike Neural Network, biological plausibility, neural networks

@booklet{EasyChair:11905,
  author    = {Sanaullah and Shamini Koravuna and Ulrich Rückert and Thorsten Jungeblut},
  title     = {Design-Space Exploration of SNN Models Using Application-Specific Multi-Core Architectures},
  howpublished = {EasyChair Preprint 11905},
  year      = {EasyChair, 2024}}