I will join the Department of Computer Science and Software Engineering at the University of Canterbury as a Lecturer in January 2027.
I am currently an Honorary Research Associate at the Artificial Intelligence Institute (Te Ipu o te Mahara), University of Waikato, where I previously worked as a Research Fellow.
Before moving to New Zealand, I was a Research Assistant Professor in the Department of Computer Science and Engineering at Shanghai Jiao Tong University (SJTU). I received my PhD in Computer Science from Nanyang Technological University (NTU) in 2020 and my Bachelor’s degree from SJTU in 2015.
Research
Biological neural networks are natural lifelong learners. They continually acquire new skills, consolidate what they learn, and adapt to ever-changing circumstances. For artificial neural networks, learning something new can unexpectedly disrupt what a model already knows, often in ways that are hard to predict.
I want to understand how intelligent systems learn and grow from incoming experience and data. What determines whether new learning builds on existing knowledge or interferes with it? When is forgetting a failure, and when is it a necessary part of adaptation? And as AI systems move from static deployment to continual adaptation, how can we keep them predictable, controllable, and beneficial to the people who interact with them?
My work brings together three closely related research areas:
-
Graceful Continual Adaptation: Understanding when and why forgetting occurs, and developing methods that allow AI systems to learn new knowledge and capabilities while selectively preserving or revising existing ones. I am particularly interested in making continual adaptation more predictable and controllable.
-
Sample-Efficient Reinforcement Learning: Developing statistically and computationally efficient algorithms that learn from interaction and feedback. My work explores policy gradient methods and decision-making under limited feedback, particularly in settings where collecting interaction data is costly.
-
Human–AI Interaction: Understanding how people interact with and adapt to intelligent systems, and designing AI systems that respond to users’ changing needs and behaviours. My work includes interdisciplinary research on adaptive learning environments, intelligent tutoring systems, and human-centred applications.
Publications
-
Agreement-gated distillation for selective forgetting in knowledge revisionNeurIPS Workshop on Autonomous Machine Learning Research (accepted), 2026
-
Rethinking memory in continual learning: Beyond a monolithic store of the pastTransactions on Machine Learning Research (TMLR), 2025 ★ J2C Certification
-
Kolmogorov-Arnold networks still catastrophically forget but differently from MLPProceedings of the 39th AAAI Conference on Artificial Intelligence (AAAI), 2025
-
Look at me, no replay! SurpriseNet: Anomaly detection inspired class incremental learningProceedings of the 32nd ACM International Conference on Information and Knowledge Management (CIKM), 4038–4042, 2023
-
Repeated augmented rehearsal: A simple but strong baseline for online continual learningAdvances in Neural Information Processing Systems 35 (NeurIPS), 2022
-
Balancing the stability-plasticity dilemma with online stability tuning for continual learningInternational Joint Conference on Neural Networks (IJCNN), 2022
-
Personalized task difficulty adaptation based on reinforcement learningUser Modeling and User-Adapted Interaction 31(4), 753–784, 2021
-
Bootstrapped policy gradient for difficulty adaptation in intelligent tutoring systemsProceedings of the 18th International Conference on Autonomous Agents and MultiAgent Systems (AAMAS), 2019
-
The influence of peer accountability on attention during gameplayComputers in Human Behavior 84, 18–28, 2018
-
Curvature-based method for determining the number of clustersInformation Sciences 415, 414–428, 2017
Selected projects
Reinforcement Learning
Sample efficient bootstrapped policy gradient for difficulty adaptation
Intelligent Tutoring Systems
Dynamic difficulty adjustment for visual memory training
Teaching
- COSC442-27S2 Natural Language Processing — University of Canterbury, Semester 2, 2027
- CS7327-033-M01 Neural Network Theory and Applications (Graduate course) — Shanghai Jiao Tong University, Spring 2021