Classical data structures vs. learned approaches
Project Leader: Renata Borovica-Gajic
Collaborators: Martin Tomko(Infrastructure Engineering)
Primary Contact: Renata Borovica-Gajic (email@example.com)
Keywords: data structures; database systems; machine learning
Disciplines: Computing and Information Systems
There has recently been a lot of excitement about a new proposal from authors at Google: to replace conventional indexing data structures like B-trees and hash maps by instead fitting a neural network to the dataset. The paper compares such learned indexes against several standard data structures and reports promising results. While learned indexes are an exciting research area, traditional data structures have been optimized for decades.
In this proposal we will look into learned indexes targeted for multidimensional data such as spatial data. We will investigate whether spatial data requires a new type of indexes, and whether such indexes can be learned by Machine Learning and deep network models.
Further information: https://arxiv.org/abs/1712.01208