Introduces Clipped Linear Lottery with linear probability scaling between thresholds, proving its worst-case regret nearly matches the lower bound for any smooth selection rule and showing improved stability on real peer-review datasets.
For completeness, Algorithm 2 outlines an exact, non-iterative implementation with runtimeO(n 2), adapted from Wang and Lu (2015)
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
-
Smooth Partial Lotteries for Stable Randomized Selection
Introduces Clipped Linear Lottery with linear probability scaling between thresholds, proving its worst-case regret nearly matches the lower bound for any smooth selection rule and showing improved stability on real peer-review datasets.