Single-pass Streaming Lower Bounds for Multi-armed Bandits Exploration with Instance-sensitive Sample Complexity.
Sepehr AssadiChen WangPublished in: NeurIPS (2022)
Keyphrases
- sample complexity
- single pass
- lower bound
- multi armed bandits
- stream mining
- upper bound
- vc dimension
- pac learning
- bandit problems
- upper and lower bounds
- data streams
- optimal solution
- objective function
- np hard
- concept class
- theoretical analysis
- multi armed bandit
- learning problems
- concept classes
- worst case
- active learning
- learning algorithm
- sample size
- statistical queries
- supervised learning
- average case
- data mining
- semi supervised learning
- regret bounds
- special case