-
Slava Sazonau
(Viachaslau Sazonau)
-
Contact me
Room 2.89, Kilburn Building
School of Computer Science
University of Manchester
Oxford Road, Manchester
M13 9PL, UKAbout me
I am a computer scientist at the University of Manchester where I hold a research associate position in the School of Computer Science.
My research interests are in the broad area of Artificial Intelligence:
knowledge representation and reasoning (Description Logics, Semantic Web, OWL, RDF)
machine learning and data mining, in particular Semantic Web applications
natural language processing and understanding, in particular knowledge-aware and ontology-based methods
I enjoy investigating avenues where knowledge and data, reasoning and learning meet together and nice applications can potentially emerge from that convergence. My PhD in semantic data mining (Link), under amazing supervision of Uli Sattler, was devoted to one of such avenues.
If you are interested in one of these topics, just drop me a line!
My research
I have investigated computational performance of Description Logic/OWL reasoners which are highly optimized implementations of algorithms for deductive reasoning over ontologies. Due to complexity of the input, their performance varies significantly across inputs. I have analysed ontological features and developed algorithms that predict reasoning performance by inspecting the input ontology.
I have done research in expressive probabilistic logics, in particular, extensions of the standard semantics of Description Logics to capture probabilistic knowledge. I have investigated probabilistic reasoning problems and their computational complexity.
I am currently investigating the problem of mining ontological knowledge (logical formulae such as “having some children implies being a father or mother”) from labelled graph data. The developed approach, which is explained in detail in my PhD thesis, can be used toautomatically build or complete a domain ontology
reveal new interesting domain knowledge
indicate modelling errors or poor data quality
You are welcome to try the tool, called DL-Miner, which is publicly available as a
web service: dlminer.io (you can use kinship as a simple input file)
plugin for the Protege editor: protege.dlminer
Publications
V. Sazonau, U. Sattler "Mining Hypotheses from Data in OWL: Advanced Evaluation and Complete Construction" 16th International Semantic Web Conference (ISWC 2017), Vienna, Austria Link
V. Sazonau "General Terminology Induction in Description Logics" PhD thesis, The University of Manchester, 2017 Link
V. Sazonau, U. Sattler, G. Brown "General Terminology Induction in OWL" 14th International Semantic Web Conference (ISWC 2015), Bethlehem, Pennsylvania, USA Link PDF
V. Sazonau, U. Sattler, G. Brown "General Terminology Induction in OWL" 12th OWL: Experiences and Directions Workshop (OWLED 2015), Bethlehem, Pennsylvania, USA Link PDF
V. Sazonau, U. Sattler "TBox Reasoning in the Probabilistic Description Logic SHIQp" 28th International Workshop on Description Logics (DL 2015), Athens, Greece, CEUR Workshop Proceedings Link PDF
V. Sazonau, U. Sattler, G. Brown "Predicting Performance of OWL Reasoners: Locally or Globally?" 14th International Conference on Principles of Knowledge Representation and Reasoning (KR 2014), Vienna, Austria, AAAI Press Link PDF
V. Sazonau, U. Sattler, G. Brown "Predicting OWL Reasoners: Locally or Globally?" 27th International Workshop on Description Logics (DL 2014), Vienna, Austria, CEUR Workshop Proceedings Link PDF
V. Sazonau "Performance Prediction of OWL Reasoners" Master's thesis, The University of Manchester, 2013 PDF
-