嵌入基于案例记忆机制的个人辅助Agent

Personal Assistant Agents with a Case-based Memory

个人辅助agent是一种能够帮助用户解决工作环境中特定任务的认知agent。为个人辅助agent设计了一种记忆机制,用于表示和组织个人辅助agent的记忆。受基于案例推理思想的启发,构建了一个记忆处理器,以对记忆进行更新和管理。在多agent平台——OMAS中实现了个人辅助agent原型系统。实验表明,具有记忆机制的个人辅助agent能更有效地执行用户任务。

Personal Assistant（PA）agents are cognitive agents capable of helping users handle tasks at their workplace.The paper concerned the design of a memory mechanism for PA agents,which leads to fairly complex cognitive agents.Inspired by the idea of Case-based Reasoning（CBR）method,a cognitive memory processor was built to monitor and update the memory of PA.A prototype of the PA agents was developed and implemented in our multi-agent platform named OMAS.Several experiments show the memory makes the PA agent more efficient in handling the tasks requested by the user.