Temporal-Difference Learning Using Distributed Error Signals Jonas Guan

It is unclear whether distributed errors alone are sufficient for synapses to make coordinated updates to learn complex, nonlinear reward-based learning tasks.