Learning to learn online with neuromodulated synaptic plasticity in spiking neural networks
This work provides a new approach for developing neuroscience-inspired online learning algorithms, which could benefit researchers in computational neuroscience and machine learning seeking brain-like models.
The authors tackled the challenge of applying neuroscience-derived learning models to achieve performance comparable to deep learning methods by training neuromodulated synaptic plasticity in Spiking Neural Networks using a learning-to-learn framework with gradient descent, enabling these models to address online learning problems effectively.
We propose that in order to harness our understanding of neuroscience toward machine learning, we must first have powerful tools for training brain-like models of learning. Although substantial progress has been made toward understanding the dynamics of learning in the brain, neuroscience-derived models of learning have yet to demonstrate the same performance capabilities as methods in deep learning such as gradient descent. Inspired by the successes of machine learning using gradient descent, we demonstrate that models of neuromodulated synaptic plasticity from neuroscience can be trained in Spiking Neural Networks (SNNs) with a framework of learning to learn through gradient descent to address challenging online learning problems. This framework opens a new path toward developing neuroscience inspired online learning algorithms.