A new design and a sharper analysis of orthogonalized regression yield optimal regret, logarithmic regret under gaps, and the first PAC and best-arm identification guarantees for semiparametric bandits.
We prove the upper bound of the theorem in four steps; afterward, we show its optimality
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
-
Experimental Design for Semiparametric Bandits
A new design and a sharper analysis of orthogonalized regression yield optimal regret, logarithmic regret under gaps, and the first PAC and best-arm identification guarantees for semiparametric bandits.