NSF AI Institute TILOS: Advancing AI and Optimization for Intelligent Systems

TILOS Phase II targets next-generation artificial intelligence (AI) research and education by focusing on physically embedded real-world intelligence. Embedded intelligence integrates AI, and even larger foundation models, with physical embodiment to allow real-time decision making, control, and interaction within complex environments. This is critical to areas such as transportation, manufacturing, healthcare, energy systems, and scientific instruments. The project builds upon its first five years to move beyond learning-enabled optimization toward a new scientific paradigm for embedded intelligent systems, where optimization is central to how AI learns, scales, coordinates, and operates in complex real-world environments. The Institute has significant broader impact activities through translation, education and workforce development, outreach, community building, and open-source release of new frameworks for AI and optimization. By leveraging its strong relationship with leading industry partners across sectors critical to the national economy, the institute further accelerates knowledge transfer to the private sector, provides interdisciplinary student training in AI applications, and expands industry partnerships for pilot deployments in domains such as robotics, communication networks, and chip design. A convergent set of foundational scientific challenges are represented in the three thrusts of the Institute’s research agenda. The thrusts target modern AI systems that (i) adapt, reason, and generalize more effectively, efficiently and robustly than current methods (Thrust 1: Foundations); (ii) act and coordinate in the physical world as teams of embodied agents (Thrust 2: Embodied Multi-Agent AI); and (iii) operate reliably and efficiently across chips and system substrates (Thrust 3: AI Hardware Substrate). Across these thrusts, the Institute investigates new theory, algorithms, and systems that unify foundational research with high-stakes application domains, helping define the next era of scalable, trustworthy, and physically grounded intelligent systems. The Institute also serves as a national nexus for leadership at the intersection of AI, optimization, and critical applications. Through workshops on cutting-edge topics, AI Futures Roundtables, Grand Challenge events, and other activities, the Institute leads as a hub where researchers, educators, industry leaders, and public stakeholders come together to push boundaries, and collectively define the next frontier of intelligent systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.