On the performance of learned data structures

A recent trend in algorithm design consists of augmenting classic data structures with machine learning models, which are better suited to reveal and exploit patterns trends the input so achieve outstanding practical improvements space occupancy time efficiency. This is especially known context indexing for big where, despite few attempts evaluating their asymptotic efficiency, theoretical results yet missing showing that learned indexes provably than indexes, such as B-tree s variants. In this paper, we present first mathematically-grounded answer problem by exploiting a link mean exit over proper stochastic process which, show, related complexity these indexes. As corollary general analysis, show plugging result (learned) PGM-index, get structure s.