An ε-agnostic algorithm for fixed-budget ε-good max-min action identification in depth-2 trees achieves misidentification probability decaying as exp(-~Θ(T/H₂(ε))).
By Lemma 8, ˆ∆max > 3 4∆(KL), hence ˆ∆y,1 < ˆ∆max, and (y,1) will not be eliminated by the single-leaf elimination rule
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
-
$\varepsilon$-Good Action Identification in Fixed-Budget Monte Carlo Tree Search
An ε-agnostic algorithm for fixed-budget ε-good max-min action identification in depth-2 trees achieves misidentification probability decaying as exp(-~Θ(T/H₂(ε))).