Abstract
Learned indices have been proposed to replace classic index structures like B-Tree with machine learning (ML) models. They require to replace both the indices and query processing algorithms currently deployed by the databases, and such a radical departure is likely to encounter challenges and obstacles. In contrast, we propose a fundamentally different way of using ML techniques to improve on the query performance of the classic R-Tree without the need of changing its structure or query processing algorithms. Specifically, we develop reinforcement learning (RL) based models to decide how to choose a subtree for insertion and how to split a node, instead of relying on hand-crafted heuristic rules as R-Tree and its variants. Experiments on real and synthetic datasets with up to 100 million spatial objects clearly show that our RL based index outperforms R-Tree and its variants.
Abstract (translated)
URL
https://arxiv.org/abs/2103.04541
