Federated Foundation Models for Recommendations. Foundation model (FM) is a machine learning term to describe the technology of developing large language models. This project aims to develop an FM-emp

Description

Federated Foundation Models for Recommendations. Foundation model (FM) is a machine learning term to describe the technology of developing large language models. This project aims to develop an FM-empowered recommendation framework with powerful modeling capacity, privacy preservation, and fine-grained personalisation. The project's outcomes can enhance existing recommendation models by leveraging the changing preferences of users and evolving tendencies with privacy preservation. The project can benefit Australian users by improving recommendation services with greater privacy protection and better user experience. Anticipated outcomes include new knowledge, algorithms, and toolkits for use in developing new service architecture in real applications, such as video and commercial goods.. Scheme: Discovery Projects. Field: 4611 - Machine Learning. Lead: A/Prof Guodong Long