Agents United: an open platform for multi-agent conversational systems.

Beinema, Tessa; Davison, Daniel; Reidsma, Dennis; Snaith, Mark

Abstract

The development of applications with intelligent virtual agents (IVA) often comes with integration of multiple complex components. In this article we present the Agents United Platform: an open source platform that researchers and developers can use as a starting point to setup their own multi-IVA applications. The new platform provides developers with a set of integrated components in a sense-remember-think-act architecture. Integrated components are a sensor framework, memory component, Topic Selection Engine, interaction manager (Flipper), two dialogue execution engines, and two behaviour realisers (ASAP and GRETA) of which the agents can seamlessly interact with each other. This article discusses the platform and its individual components. It also highlights some of the novelties that arise from the integration of components and elaborates on directions for future work.

Citation

BEINEMA, T., DAVISON, D., REIDSMA, D. et al. 2021. Agents United: an open platform for multi-agent conversational systems. In Proceedings of 21st ACM (Association for Computing Machinery) Intelligent virtual agents international conference 2021 (IVA '21), 14-17 September 2021, [virtual conference]. New York: ACM [online], pages 17-24. Available from: https://doi.org/10.1145/3472306.3478352

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 21st ACM (Association for Computing Machinery) Intelligent virtual agents international conference 2021 (IVA '21)
• Start Date: Sep 14, 2021
• End Date: Sep 17, 2021
• Acceptance Date: Jun 15, 2021
• Online Publication Date: Sep 14, 2021
• Publication Date: Sep 14, 2021
• Deposit Date: Jan 11, 2022
• Publicly Available Date: Jan 11, 2022
• Publisher: Association for Computing Machinery (ACM)
• Peer Reviewed: Peer Reviewed
• Pages: 17-24
• Book Title: Proceedings of 21st ACM (Association for Computing Machinery) Intelligent virtual agents international conference (IVA '21)
• ISBN: 9781450386197
• DOI: https://doi.org/10.1145/3472306.3478352
• Keywords: Open source platform; Multi-agent system; Social conversational agents; Coaching
• Public URL: https://rgu-repository.worktribe.com/output/1569558