Correctness is its own reward: bootstrapping error signals in self-guided reinforcement learning
Ziyi Gong, Fabiola Duarte, RICHARD D. MOONEY, John Pearson
This paper models how juvenile zebra finches learn song through self-guided reinforcement learning by developing a neural circuit that generates internal error signals via predictive cancellation.
bioRxiv · Jul 18, 2025 · PreprintNeuroscience