Introduces self-degrading Markovian bandits and UCB-NOM algorithm achieving nearly logarithmic regret without prior knowledge and O(log T) with bias bounds, with bounds independent of state count.
Hippolyte Bourel, Odalric Maillard, and Mohammad Sadegh Talebi
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Learning in Markovian bandits with non-observable states and constrained decision epochs
Introduces self-degrading Markovian bandits and UCB-NOM algorithm achieving nearly logarithmic regret without prior knowledge and O(log T) with bias bounds, with bounds independent of state count.