Modern Query Evaluation Algorithms
When: Thursday, October 13, 2022, 3pm (CEST).
Where: Zoom online, check the address in the Google Calendar Event.
Topic: Modern Query Evaluation Algorithms.
Speaker: Nadezda (Nadia) Knorozova, Ph.D. student at the University of Zurich.
Abstract
I will talk about the recent advances in the query evaluation algorithms for join and aggregate queries, focusing on the theoretical research that is implemented in commercial database systems.
Since 2011 there has been a new understanding of the complexity of join queries. This led to an array of new theoretical results and eventually to more efficient algorithms. I will focus on state-of-the-art methods that have been successfully absorbed by commercial database systems. In particular, I will present the Leapfrog Triejoin algorithm for evaluation of join queries, and the FAQ framework for evaluation of aggregate queries that is based on decomposition of query hypergraphs.
Short Bio
Nadezda Knorozova, or simply Nadia, is a Ph.D. student at the University of Zurich, with the background in relational databases with both academic and industry experience. She worked on the commercial Datalog engine called Logicblox, covering many aspects of the system: in-database machine learning, which implements machine learning algorithms that are able to exploit relational structure of the data; the query optimizer, a search algorithm for finding the best join order; join and aggregate query evaluation.
Nadezda received her bachelor degree in Computer Science from the University of Minnesota. There, in collaboration with the database group, she has developed a novel decomposition-based algorithm for computing Voronoi diagrams. She received her Masters degree in Computer Science from the University of Oxford, with a thesis on factorized representations of disequality joins.