Open Problem: Model Selection for Contextual Bandits
arXiv:2006.10940v119 citations
Originality Synthesis-oriented
AI Analysis
This addresses a theoretical gap in contextual bandit algorithms for researchers, but it is an open problem with no incremental progress reported.
The paper investigates whether model selection guarantees, which adapt to the best hypothesis class complexity in statistical learning, can be extended to contextual bandit learning, but does not provide specific results or numbers.
In statistical learning, algorithms for model selection allow the learner to adapt to the complexity of the best hypothesis class in a sequence. We ask whether similar guarantees are possible for contextual bandit learning.