Mobile scheduling assistant

Abstract

Although, in recent years many new calendaring tools have become available, most individuals still spend a great amount of time, personally performing scheduling tasks. Apart from that, recent developments in the mobile device market have made it imperative for any calendaring tool to be also accessible from a mobile device. Our goal, is to add a minimal amount of intelligence to our electronic devices, such that we can help to reduce the user's workload. This has led us to develop a scheduling assistant that can perform scheduling tasks on behalf of the user. The resulting assistant is a learning user agent that allows the user to retain control at times and relinquish it at others. We propose an assistant that may be instructed both from a mobile device as well as a standard PC. The main elements in our approach are, the use of an existing agent framework which promotes extensibility and facilitates peer-to-peer communication, learning techniques to learn patterns in the user's behaviour and user guidance to enable learning by direct instruction. With frequent use, the system begins to improve its model of the user until eventually user guidance will no longer be necessary.