Abstract

In a multi-agent system, the actions of one or more agents may lead to undesirable results that may affect the entire system. As a result, it is important to have governance and regulatory controls in place. Policies are commonly used to establish constraints on the actions permitted or prohibited within a system. However, it is challenging for policy authors to manually formulate actionable policies in complex multi-agent systems. In this talk, we will discuss current research in the field and introduce preliminary work on a framework that applies large language models (LLMs) to policy generation and management, simplifying the work of humans in the loop. LLMs provide powerful capabilities in natural language processing that are currently being explored for policy generation, management, and governance of distributed adaptive systems.

About the Speaker

Marco M. Carvalho is a Professor at the Florida Institute of Technology, in Melbourne, FL/USA. He graduated in Mechanical Engineering from the University of Brasilia (UnB, Brazil), where he also completed his M.Sc. in Mechanical Engineering with a specialization in dynamic systems. Dr. Carvalho also holds an M.Sc. in Computer Science from the University of West Florida and a Ph.D. in Computer Science from Tulane University, with specialization in Machine Learning and Data Mining.