For fixed-budget change point identification in piecewise constant bandits, the paper proves near-matching upper and lower bounds on error probability and gives an adaptive algorithm that is near-optimal in both small and large budget regimes.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Fixed-Budget Change Point Identification in Piecewise Constant Bandits
For fixed-budget change point identification in piecewise constant bandits, the paper proves near-matching upper and lower bounds on error probability and gives an adaptive algorithm that is near-optimal in both small and large budget regimes.