Fast Factorized Learning: Powered by In-Memory Database Systems

TL;DR

This paper demonstrates that in-memory database systems significantly accelerate factorized learning models, achieving up to 100 times faster training compared to traditional disk-based systems, by pre-computing shared cofactors.

Contribution

It provides the first open-source implementation of in-database factorized learning, benchmarking its performance on PostgreSQL and HyPer, showing substantial speed improvements.

Findings

70% faster than non-factorized learning on in-memory systems

100x faster than disk-based database systems

Modern in-memory databases can significantly speed up machine learning pipelines

Abstract

Learning models over factorized joins avoids redundant computations by identifying and pre-computing shared cofactors. Previous work has investigated the performance gain when computing cofactors on traditional disk-based database systems. Due to the absence of published code, the experiments could not be reproduced on in-memory database systems. This work describes the implementation when using cofactors for in-database factorized learning. We benchmark our open-source implementation for learning linear regression on factorized joins with PostgreSQL -- as a disk-based database system -- and HyPer -- as an in-memory engine. The evaluation shows a performance gain of factorized learning on in-memory database systems by 70\% to non-factorized learning and by a factor of 100 compared to disk-based database systems. Thus, modern database engines can contribute to the machine learning pipeline by pre-computing aggregates prior to data extraction to accelerate training.