IR, NLP, Deep Learning
Title of Thesis
Personalizing Query Auto-Completion system using a re-ranker
Description of Thesis
We are going to design a reranker system that considers a K-best list including candidate queries and their scores as inputs and outputs a reranked list. The inputs can be provided from a basic QAC system such as MPC or a personalized language model for QAC. The reranked list will be generated using some features such as time and location by applying a classifier or a learning-to-rank algorithm.