FST Graduate Research Topic

Research Topic: Reinforcement learning for natural language generation
Suggested By: Dunwei (Grant) Wen
Last Update: September 2nd, 2009

This project aims to use reinforcement learning, a machine learning technique, to enhance natural language generation (NLG), a branch of natural language processing (NLP). NLG can be used in spoken dialogue systems, intelligent tutoring systems, call centers and computer games for facilitating human-computer interaction and conversation. The goal of this research is to explore certain reinforcement learning strategies for improving natural language generation, build a natural language generation system in the context of a chosen application scenario such as second language learning and computer games, and to investigate and implement a spoken dialog system capable of either speech or textual human-computer conversation for providing certain services such as consulting and advising services in a certain domain. This project will explore some issues such as user modeling, semantics and intention analysis, and dialogue management structure etc. based on the state-of-the-art research and application of natural language processing. A protocol system will be developed for a business, entertainment or e-learning application. The graduate students who want to pursue the research need to be familiar with AI especially machine learning, natural language processing, and uncertainty reasoning techniques. This research topic is for Project or Thesis.