Dana Nau Publishes New AI Book on Acting, Planning and Learning

Artificial intelligence systems are becoming increasingly capable of making decisions, adapting to new information and performing complex actions. Behind these abilities are fundamental processes such as planning, acting and learning, functions that researchers are working to better understand and integrate. A new book by University of Maryland Professor Emeritus Dana Nau and his colleagues examines these processes and the challenges involved in uniting them.

The release of Acting, Planning, and Learning builds on this exploration, examining how artificial intelligence systems can better emulate these essential cognitive functions. Co-authored by Nau with Malik Ghallab of LAAS-CNRS at Toulouse University in France and Paolo Traverso of Fondazione Bruno Kessler in Italy, the book highlights capabilities that are foundational for AI “actors,” including physical robots, virtual agents in simulated environments and other autonomous systems.

“AI actors require essential cognitive functions without which intelligence is hardly conceivable, and this book is about three of them,” Nau said.

Acting, Planning, and Learning integrates state-of-the-art techniques and identifies open research challenges in AI, robotics and cognitive systems. The final manuscript is available for free online, along with lecture slides designed to support teaching and research.

This fall, the book and accompanying slides will be used in CMSC 722: AI Planning, a computer science graduate course that is also popular with ISR and with students from the Institute for Systems Research and the Maryland Robotics Center. The course, historically taught by Nau, will now be led by Research Scientist Mark “Mak” Roberts.

Nau’s decades of work in automated planning and decision-making continue to shape how researchers approach intelligence in both theory and practice. His latest publication offers a guide for scholars, educators and practitioners working at the intersection of artificial intelligence, systems engineering and autonomy.

—Article adapted from a news release by ISR