Multiagent Systems for Educational Transformation: A Swarm of Intelligent Personal Assistants
Multiagent Systems for Educational Transformation: A Swarm of Intelligent Personal Assistants
Sergio Arjona-Giner and Faraón Llorens-Largo
University of Alicante, Alicante, Spain
Technological Ecosystems for Enhancing Multiculturality (TEEM 2024)
23-25 October 2024
Alicante, Spain
2024.teemconference.eu
Proceedings
Lecture Notes in Educational Technology
Abstract.
The arrival of LLMs in November 2022, marked by the release of the
GPT-3.5 model, represents a milestone in the field of AI. These advanced mod-
els have demonstrated an exceptional ability to comprehend and generate hu-
man-like text, among other skills. The versatility and proficiency of LLMs have
paved the way for the development of various applications, including intelligent
agents. When these agents are combined synergistically in a MAS, they utilize
the trained or pre-trained LLMs as their core AI engine. In the educational sec-
tor, this AI-driven transformation has reinforced the concept of Smart Learning,
leveraging advanced technologies to create interactive, personalized, and effi-
cient learning experiences. One tangible manifestation of Smart Learning is the
development of SLApps which harness AI for robust educational support. In
this paper, we explore the introduction of a SLApp that employs a swarm of in-
telligent agents coordinated within a MAS. By leveraging the capabilities of
LLMs and the coordinated efforts of intelligent agents, the SLApp aims to im-
prove the educational process. Key contributions include the preliminary archi-
tecture of the SLApp, integration with the LTI 1.3 standard for compatibility
with LMSs and implementation of RAG technology to minimize hallucinations
and enhance knowledge sources.
Keywords:
Multiagent System, RAG, Smart Learning.