Lifelong Paradigms for Versatile, Robust and Agile Recommender Systems (2025-2028)
Abstract
This project aims to revolutionise the traditional static recommendation with novel lifelong paradigms. This project expects to develop the next generation of recommender systems to empower decision making in a fast-changing world. Expected outcomes include a recommendation platform with versatile models to handle heterogeneous data, robust test-time operations to enhance deployment effectiveness, and agile update strategies for efficient model evolution. The success of this project should significantly benefit the society and economy via developing cutting-edge techniques to reduce the energy cost and protect the privacy for Australia¿TM)s Science and Research Priorities sectors, such as cybersecurity, energy, and environmental change.