Tag
The paper presents a dynamical mean-field theory for learning-induced transitions in recurrent neural networks, showing how feedback-driven learning shifts network dynamics from chaos to stability.