PAC-Bayesian Lifelong Learning For Multi-Armed Bandits.
Hamish FlynnDavid ReebMelih KandemirJan PetersPublished in: CoRR (2022)
Keyphrases
- lifelong learning
- multi armed bandits
- pac bayesian
- distribution free
- bandit problems
- learning processes
- data dependent
- information and communication technologies
- e learning
- m learning
- professional development
- generalization bounds
- rademacher complexity
- error bounds
- learning community
- online course
- multi armed bandit
- supervised learning
- decision making
- decision problems
- context aware
- support vector machine