Emotion-driven learning agent for setting rich presence in mobile telephony

Presence or personal status information is going to be an integral part of human life in the near future. With the possibility of personalizing user preferences in a fine grained way, mobile presence will appeal to most users. Among all the advantages, one of the most annoying problems is to set the presence status manually each time. This paper discusses the development of an intelligent agent based presence client that will learn and make decisions on behalf of the user about his or her presence status. The decision is emotion driven and the learning depends on real world experience. The proposed system utilizes a neural network (NN) based emotion-driven agent to learn user preferences. As a NN learning algorithm, two approaches based on Differential Evolution and Reinforcement have been proposed, of which either one can be used. Rich presence status is set using a scripting language named Call Processing Language; and SIP is used for publishing the presence to others.