PAC-Bayesian lifelong learning for multi-armed bandits.
Hamish FlynnDavid ReebMelih KandemirJan PetersPublished in: Data Min. Knowl. Discov. (2022)
Keyphrases
- lifelong learning
- multi armed bandits
- pac bayesian
- distribution free
- learning processes
- bandit problems
- information and communication technologies
- e learning
- data dependent
- multi armed bandit
- rademacher complexity
- professional development
- error bounds
- generalization bounds
- m learning
- learning community
- online course
- upper bound
- probability distribution
- mobile devices
- learning environment
- decision making