Tag
This paper introduces the Level-Constrained-Littlestone-Littlestone (LCLL) tree to characterize learnability in universal transductive online classification with possibly unbounded label spaces, proving that optimal mistake rates are either bounded or logarithmic.