Agentic Discovery: Closing the Loop With Cooperative Agents

Download PDF
Open Webpage

J. Gregory Pauloski, Kyle Chard, Ian T. Foster

Published: 2025, Last Modified: 25 Jan 2026Computer 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: As data-driven methods, artificial intelligence, and automated workflows accelerate scientific tasks, we see the rate of discovery increasingly limited by human decision-making tasks such as setting objectives, generating hypotheses, and designing experiments. Cooperative agents can augment the role of humans and enable autonomous discovery.