Temporal-Difference Learning Using Distributed Error Signals
This addresses a computational challenge in neuroscience for understanding biological learning mechanisms, but is incremental as it adapts existing deep Q-learning with a novel error distribution approach.
The paper tackled the problem of whether distributed error signals, like dopamine in the brain, can enable complex reinforcement learning without explicit credit assignment, and showed that their Artificial Dopamine algorithm achieves performance comparable to backpropagation-based methods on tasks like MinAtar and DeepMind Control Suite.
A computational problem in biological reward-based learning is how credit assignment is performed in the nucleus accumbens (NAc). Much research suggests that NAc dopamine encodes temporal-difference (TD) errors for learning value predictions. However, dopamine is synchronously distributed in regionally homogeneous concentrations, which does not support explicit credit assignment (like used by backpropagation). It is unclear whether distributed errors alone are sufficient for synapses to make coordinated updates to learn complex, nonlinear reward-based learning tasks. We design a new deep Q-learning algorithm, Artificial Dopamine, to computationally demonstrate that synchronously distributed, per-layer TD errors may be sufficient to learn surprisingly complex RL tasks. We empirically evaluate our algorithm on MinAtar, the DeepMind Control Suite, and classic control tasks, and show it often achieves comparable performance to deep RL algorithms that use backpropagation.