Posted: 15th March 2016

The Vienna Center for Logic and Algorithms (VCLA) has awarded Oxford Computer Science student Maximilian Schleich, the Honourable Mention for Outstanding International Master Thesis. Maximilian has been invited to Vienna this April to get the award and present his work.

Maximilian's thesis entitled ‘Learning Regression Models over Factorised Joins’ bridges databases and machine learning. The key contribution of his project is the connection between the computational complexity of linear regression, which is an ubiquitous machine learning task, and fundamental notions of tractability such as hypertree decompositions of database joins, which are well-established in logic, databases, constraint satisfaction, and probabilistic graphical models. In practice, his approach can be up to three orders of magnitude faster than state-of-the-art machine learning platforms, such as R, Python StatsModels, and MADlib, on real datasets.

His work led to a publication in ACM SIGMOD 2016, which is the ACM flagship conference in database and data management research. This work has been previously awarded the 2015 Hoare Prize for best MSc project in Computer Science at Oxford.

http://www.cs.ox.ac.uk/projects/FDB/

http://www.cs.ox.ac.uk/people/maximilian.schleich/