ProfessorGill Dobbie
Professor
Computer Science
Orcid identifier0000-0001-7245-0367
- ProfessorComputer Science
- +6493737599 Ext.83949 (Work)
- SCIENCE CENTRE 303 - Bldg 303, 38 PRINCES ST, AUCKLAND CENTRAL, AUCKLAND, 1010, New Zealand
RESEARCH INTERESTS
I have a wide range of research interests across artificial intelligence, data management, and software engineering. My main areas of interest pertain to machine learning, knowledge representation, ethics, information retrieval and database systems.
I have extensive experience in artificial intelligence and its application in healthcare and the environment, with specific current interests in data stream mining, concept drift detection and adversarial attacks. We have created efficient and effective algorithms that detect when concepts have drifted in a data stream and identify the type of the drift. When a concept drift is detected, the model must be adapted to remain accurate. In this situation, we are describing when it is cost effective to transfer information from one model to another. We're currently also describing algorithms that detect and defend against various adversarial attacks, such as data poisoning and member inference attacks.
Previously my research focussed on the foundations of database systems, defining logical models for various kinds of database systems, and reasoning about the correctness of algorithms in that setting. With colleagues at the National University of Singapore, I defined a data model for semistructured data (called ORA-SS), providing a language independent description of the data. With this group I used the ORA-SS data model to define a normal form for ORA-SS schema, defined valid views for semistructured databases, and described a storage structure for semistructured databases using object relational databases.
I have a track record of attracting over $45m of research funding from the Marsden Fund, government agencies and industry in New Zealand and internationally. I led the development of the bid for the Precision Driven Health Joint Research project, which is a collaboration between industry, district health boards and the University, developing solutions for healthcare through the translation of research. The partnership has been a success hosting around 100 projects and receiving the MinterEllisonRuddWatts Research & Business Partnership Award in recognition of the success of the collaboration.
I have extensive experience in artificial intelligence and its application in healthcare and the environment, with specific current interests in data stream mining, concept drift detection and adversarial attacks. We have created efficient and effective algorithms that detect when concepts have drifted in a data stream and identify the type of the drift. When a concept drift is detected, the model must be adapted to remain accurate. In this situation, we are describing when it is cost effective to transfer information from one model to another. We're currently also describing algorithms that detect and defend against various adversarial attacks, such as data poisoning and member inference attacks.
Previously my research focussed on the foundations of database systems, defining logical models for various kinds of database systems, and reasoning about the correctness of algorithms in that setting. With colleagues at the National University of Singapore, I defined a data model for semistructured data (called ORA-SS), providing a language independent description of the data. With this group I used the ORA-SS data model to define a normal form for ORA-SS schema, defined valid views for semistructured databases, and described a storage structure for semistructured databases using object relational databases.
I have a track record of attracting over $45m of research funding from the Marsden Fund, government agencies and industry in New Zealand and internationally. I led the development of the bid for the Precision Driven Health Joint Research project, which is a collaboration between industry, district health boards and the University, developing solutions for healthcare through the translation of research. The partnership has been a success hosting around 100 projects and receiving the MinterEllisonRuddWatts Research & Business Partnership Award in recognition of the success of the collaboration.