me

Contact


Visiting Address

De Boelelaan 1105
Main Building level 13
(closed flex-working space: call before visit)
1081 HV Amsterdam, The Netherlands
View in google maps

Filip Ilievski

Since March 2015 I am a PhD student at Vrije Universiteit Amsterdam under supervision of researchers from the Computational Lexicology and Terminology Lab (Prof Dr. Piek Vossen and Dr. Marieke van Erp) and from the Knowledge Representation & Reasoning group (Prof. Dr. Stefan Schlobach and Prof. Dr. Frank van Harmelen).

I teach machines to disambiguate long tail identities mentioned in text, by applying extensive contextual knowledge and reasoning techniques. I have studied system performance in relation to evaluation data properties, and managed to distinguish easy 'head' identity cases from difficult 'tail' cases. After observing that long tail identity is not only difficult for NLP systems, but also underrepresented in evaluation data, I co-organized SemEval-2018 task 5, which deliberately targetted long tail entities and events. I am the main developer of LOTUS, the largest Linked Open Data text index available. During the period March-September 2017, I visited the Language Technologies Institute at Carnegie Mellon University, and I investigated a new cognitive task on profiling, under supervision of Prof. Eduard Hovy.

Prior to my PhD, I completed Artificial Intelligence Master at VU Amsterdam with a cum laude honor. My education has been funded by several scholarships, including the VU Fellowship Program and the Erasmus Mundus Euroweb Program. I am proficient in a number of programming languages, including Python, Javascript/Node.js, and C#.

View my profile on Google Scholar   View Filip Ilievski's profile on LinkedIn View my profile on Github