Distributed Bandit Learning: How Much Communication is Needed to Achieve (Near) Optimal Regret.
Yuanhao WangJiachen HuXiaoyu ChenLiwei WangPublished in: CoRR (2019)
Keyphrases
- online learning
- distributed learning
- learning algorithm
- reinforcement learning
- learning process
- distributed systems
- cooperative
- neural network
- learning systems
- communication networks
- spatially distributed
- distributed computation
- inductive inference
- communication cost
- linear regression
- information sharing
- knowledge acquisition
- peer to peer
- worst case
- multi class
- upper bound
- prior knowledge