Learned Cardinalities: Estimating Correlated Joins with Deep Learning Article Swipe

PDF

Related Concepts

Joins Computer science Tuple Cardinality (data modeling) Query optimization Predicate (mathematical logic) Set (abstract data type) Artificial intelligence Data mining Deep learning Semantics (computer science) Theoretical computer science Machine learning Mathematics Programming language Discrete mathematics

Andreas Kipf , Thomas Kipf , Bernhard Radke , Viktor Leis , Peter Boncz , Alfons Kemper ·

YOU? · · 2018 · Open Access · · DOI: https://doi.org/10.48550/arxiv.1809.00677 · OA: W2890276152

We describe a new deep learning approach to cardinality estimation. MSCN is a multi-set convolutional network, tailored to representing relational query plans, that employs set semantics to capture query features and true cardinalities. MSCN builds on sampling-based estimation, addressing its weaknesses when no sampled tuples qualify a predicate, and in capturing join-crossing correlations. Our evaluation of MSCN using a real-world dataset shows that deep learning significantly enhances the quality of cardinality estimation, which is the core problem in query optimization.