Optimizing queries in probabilistic graph databases

Details

Supervisors

Nijssen, Siegfried

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil en informatique

Abstract

Traditional databases reflect deterministic models, what is stored inside them is considered as always true. However we sometimes need to add uncertainties to those databases. At a time where automation and machine learning is centres of interest, probabilistic databases are useful to model automatic translations, natural language comprehension, etc. Unfortunately, computing queries in a probabilistic database is known to be a hard problem. This is even more the case for probabilistic graphs. Therefore this thesis aims at exploring whether we can develop improved algorithms for computing the probability of reachability. Thus several alternatives were implemented. Another goal of this thesis will then be to compare those different alternatives.