The Task and Workslip Scheme for Personal Agent Systems

Tyson Kendon, Jorg Denzinger
2020 2020 IEEE International Conference on Humanized Computing and Communication with Artificial Intelligence (HCCAI)  
The Task and Workslip Scheme for Personal Agent Systems is a collection of concepts enabling the intelligent communication of small agents, and a set of philosophical guidelines to aid in the design of intelligent systems to assist people in day-to-day tasks. The intention behind the scheme is to empower people to take control of their own computing power to achieve the kind of automatic personal assistance that otherwise can only be achieved with corporate computing. The scheme addresses the
more » ... sue of how to provide people with these tools in three ways. it provides a theoretical framework for the design of small pieces of software, Agents, which communicate using usefully structured information, Workslips, and can combine to form helpful systems for people, Personal Agent Systems. It provides a set of philosophical guidelines which suggest particular design choices developers should make while developing these Agents. Finally it provides a reference implementation, TWPARI-J, which can be used to develop these Agents and provides practical examples of the philosophical guidelines. To demonstrate the effectiveness of TWPARI-J and the applicability of the philosophical guidelines and conceptual framework, three example systems are presented which allow people to: adjust their alarm clocks according to the weather, plan their shopping according to sales prices of local stores, and to get assistance monitoring the state of their network in an intelligent fashion. For each system particular instantiations are provided where the system is able to help a particular person. The scheme provides a simple system to create personal assistant system, often without needing to write new software, and simplifies Agent development. The scheme allows people to use the system and develop applications with a maximum of flexibility using a minimum of work. My thanks first go to Professor Jörg Denzinger, who has been my guide, sounding board, sanity checker, and reader of endless chapters and fixes.
doi:10.1109/hccai49649.2020.00012 fatcat:3b2jogwtanbkba6xhwhitldqly