Reinforcement learning with a network of spiking agents
This work addresses reinforcement learning problems for AI researchers by offering a biologically-inspired approach, though it appears incremental as it builds on existing neuroscientific theory and GLM methods.
The authors tackled reinforcement learning tasks by proposing a multi-agent framework using spiking neurons in a generalized linear model formulation, showing it can learn complex action representations and reduce variance in learning updates through brain-inspired modularity and population coding.
Neuroscientific theory suggests that dopaminergic neurons broadcast global reward prediction errors to large areas of the brain influencing the synaptic plasticity of the neurons in those regions. We build on this theory to propose a multi-agent learning framework with spiking neurons in the generalized linear model (GLM) formulation as agents, to solve reinforcement learning (RL) tasks. We show that a network of GLM spiking agents connected in a hierarchical fashion, where each spiking agent modulates its firing policy based on local information and a global prediction error, can learn complex action representations to solve RL tasks. We further show how leveraging principles of modularity and population coding inspired from the brain can help reduce variance in the learning updates making it a viable optimization technique.