Gillian Dobbie

Gillian’s collaborative research tackles software quality before implementation by automating the diagnosis and repair of design models. The approach is intended for developers working with these models; published experiments show repairs can remove faults while preserving model quality. The continual-learning review and shared research resources also support researchers studying how learning systems can adapt without losing previously acquired knowledge.

Gillian Dobbie is Professor of Computer Science at the University of Auckland, pursuing research across data management, machine learning, logic and software engineering to address practical, real-world problems. Leadership within the wider scientific community includes serving as Chair of the Marsden Fund Council.

Alongside Cheng-Hao Cai and Jing Sun, Gillian co-authored the 2025 paper ‘The automation of design model repair’. The study applies model checking to identify design faults, then generates automated modifications to resolve unreachable goals and violated invariants. Experimental results demonstrated that the technique eliminates defects while preserving model quality, providing a practical mechanism to catch design errors before code is written.

Gillian also co-authored a 2026 systematic review examining continual learning, accompanied by an open-source repository sharing reviewed literature and resources to support reproducibility across the research community.