Optimal Protocols for Continual Learning via Statistical Physics and Control Theory

Open in new window