Enhancing Intelligence of Personal Assistant Agent Using Memory Mechanism

A Personal Assistant (PA) agent is a software agent capable of helping people to handle tasks in their workplace. The paper proposes a memory mechanism for personal assistant agents in order to enhance the agent intelligence while working with the user or other agents. A brief state of the art concerning personal assistant agents is presented first and the necessity of having a memory mechanism is explained later. In contrast to some previous work, this paper attempts to construct a memory model based on a model of episodic memory. Furthermore, a memory mechanism is designed for personal assistant agents in a multi-agent platform. Finally, future work is considered to achieve this objective in a research and development (R&D) context.