Towards A More Holistic Approach in Entity Resolution with Active Learning Algorithms

Abstract

We researched on solving entity resolution problems using active learning approaches, and focused on general-based methods. We did a literature review and presented the general pipeline for solving ER tasks with AL approaches. We critically reviewed the existing literature to thematise existing approaches in order to identify three exemplars for evaluation. We critically reflected on the findings of the exemplars and synthesised a new method, CombinedSEL. The method is critically evaluated both theoretically and empirically. It was found to outperform previous approaches. Limitations and future works were discussed as well.

Ge Wang

DPhil (Ph.D.) student

I’m a Dphil student in the Department of Computer Science at University of Oxford. My research investigates the algorithmic impact on families and children, and what that means for their long-term development. I’m keen to explore the potential for designing more age-appropriate AI for families, as well as building more ethical web and data architecture for them. My research takes a human-centric approach, and focuses on understanding users' needs in order to design technological prototypes that are of real impact on today’s society.