Zeon Trevor Fernando

1) Deep Learning for Information Retrieval using supervised and semi-supervised approaches. 2) Reproducibility and deployment of existing neural ranking models using standard ad-hoc newswire test collections. 3) Interpretability approaches to better understand the relevance decisions made by neural ranking models using both model-introspective and model-agnostic approaches.

Worked as a full-stack developer on LearnWeb: An educational web platform that supports collaborative search and sharing of open-access multimedia resources. Responsibilities: 1) Developed a search history that visualizes a knowledge graph containing important entities and their semantic relationships discovered during a search session. Accepted at SIGIR 2018. 2) Added various new features that improved user experience (organization of private resources, user learning activities automatic evaluation, etc.)

Information Retrieval, Deep Learning, Large Scale Data Mining, Web Science, User Personalization and Modeling, Efficient Algorithms

Algorithm Design and Analysis, Applied Probability, Advanced DBMS, Graph Theory, Software Engineering