Characterization of Generalizability of Spike Timing Dependent Plasticity trained Spiking Neural Networks
This work addresses the generalizability challenge in neuro-inspired machine learning for researchers in computational neuroscience and AI, but it is incremental as it builds on existing STDP methods.
The paper tackled the problem of understanding and improving the generalizability of Spiking Neural Networks trained with Spike Timing Dependent Plasticity by analyzing learning trajectories using Hausdorff dimension, and developed a Bayesian optimization approach to optimize hyper-parameters for enhanced generalizability.
A Spiking Neural Network (SNN) is trained with Spike Timing Dependent Plasticity (STDP), which is a neuro-inspired unsupervised learning method for various machine learning applications. This paper studies the generalizability properties of the STDP learning processes using the Hausdorff dimension of the trajectories of the learning algorithm. The paper analyzes the effects of STDP learning models and associated hyper-parameters on the generalizability properties of an SNN. The analysis is used to develop a Bayesian optimization approach to optimize the hyper-parameters for an STDP model for improving the generalizability properties of an SNN.