Scalable regret for learning to control network-coupled subsystems with unknown dynamics.
Sagar SudhakaraAditya MahajanAshutosh NayyarYi OuyangPublished in: CoRR (2021)
Keyphrases
- learning algorithm
- learning systems
- prior knowledge
- learning process
- active learning
- online learning
- learning problems
- adaptive control
- connectionist networks
- recurrent networks
- control system
- supervised learning
- loss function
- dynamic model
- lower bound
- reinforcement learning
- neural network
- motor control
- spiking neural networks
- partially observed