Basser Seminar Series

The Intelligent-Electronic Mail Sorter

Dr Eric McCreath
Australian National University

Wednesday 10 May 2006, 4-5pm

Basser Conference Room (Madsen Building, Room G92)



The first part of this presentation aims to bring together research that relates to the email project we (Judy Kay, Liz Crawford and myself) have been involved in over the last 5 years. Sorting email messages into folders is an important task for a vast number of computer users. The iems(Intelligent-Electronic Mail Sorter) project is focused on improving the way users interact with email managers. In this presentation I will discuss how combining learnt and hand crafted rules can improve the overall performance of the system. Also I will present a composite rule learner that classifies mail by combining an instance based approach with an approach that constructs a general explicit description. We show that this approach improves classification erformance. Furthermore, this combined approach produces understandable and concise classification rules that users can scrutinize allowing them to maintain a sense of control.

The second part of this presentation will overview the new email management system that is being developed and released. I will present research directions that will be considered. In particular we wish to consider how other work-flow activities may be combined into an email manager. An example of this is combining a "todo list" into an email manager. This would enable associations between messages and todo items and hence provide a uniform interface for handling this information. Another aspect I wish to explore on this new email manager is learning when messages are important then presenting these important messages to the user.

 Speaker's biography


Eric McCreath completed his Ph.D. degree in 1999 from the University of New South Wales. This was on research involving Inductive Logic Programming(ILP) which is a sub-field of Machine Learning. As part of his PhD thesis Dr McCreath derived a Bayesian heuristic for finding the most probable hypothesis in a very general framework that allows noisy data and fixed example size. He joined the Basser Department of Computer Science(now the School of Information Technologies) at Sydney University in 1999 and then in 2001 he joined the Department of Computer Science at the Australian National University. Dr McCreath currently holds a lecturing position at the ANU. Recently Dr McCreath has collaborated with Judy Kay from the University of Sydney investigating the application of machine learning techniques in categorizing email within a user's inbox. More recently Dr McCreath has been supervising Robert Bridle during his PhD and together they have investigated novel approach for applying machine learning to improve mobile phone interfaces.