Elizabeth Ondula and Bhaskar Krishnamachari
This research involves designing and implementing a simulation environment that incorporates variants of stochastic epidemic models to enable development, ensure robustness and effectiveness of RL-driven policies across diverse settings. It also addresses the management of inherent prediction uncertainties in RL to facilitate reliable decision-making. This research also considers challenges of maintaining fairness in strategy formulation, given the constraints and variability in space resources, alongside the development of equitable policies
Daniele Orner, Elizabeth Ondula, Nick Mumero, Casandra Rusti, Richa Goyal
In the exploration of Large Language Model (LLM)-based multi-agent systems, there emerges a need to understand and interpret the dynamics of agent interactions and their foundational beliefs, particularly when designed to emulate diverse roles and perspectives or to engage in debates. At present, there are no unified solutions that can systematically interpret and analyze the beliefs and interactions of these agents. This study introduces Sentimental Agents , a framework designed to support decision-making by providing multiple perspectives on a topic. Agents within this framework are endowed with a mental model of self, articulated in natural language. We have integrated sentiment analysis with a non-Bayesian updating mechanism to interpret and analyze the agents' beliefs and interactions systematically. A collective viewpoint is achieved when the update is marginal, enabling human observers to understand the intricacies of agents' stances and interactions within the system.
Daniel D'Souza and Elizabeth Ondula
Traditional reinforcement learning (RL) methods, when applied in isolation to such contexts, often grapple with the issue of generalization. Thus, our formulation integrates graph neural networks (GNNs) with RL. This aims to leverage GNN's capability to generalize and transfer knowledge across different scenarios, ensuring that a multi-robot system can adapt better to unforeseen situations