Thesis Title
Enhancing Personalized Search with Retrieval-Augmented Generation for Dynamic Prompt Expansion

Research Description
This thesis explores the use of Retrieval-Augmented Generation (RAG) to improve personalized search by dynamically expanding user queries with relevant contextual information. By integrating large language models with information retrieval, the system aims to enhance search relevance and user satisfaction. The research involves developing a retrieval system optimized for personalization, incorporating privacy-preserving techniques, and evaluating the approach using standard IR metrics to measure improvements in relevance, personalization, and privacy over traditional search methods.

Supervisors
Professor Mark Sanderson, RMIT University