Roman Garnett

University of Bonn, Washington University in St. Louis

Papers

Total Citations

H-Index

About

Roman Garnett is a researcher whose work spans the intersection of machine learning, robotics, and autonomous decision-making, with particular emphasis on active learning, probabilistic methods, and optimization under uncertainty. His research addresses some of the most practically significant challenges in intelligent systems, including how autonomous agents can efficiently search for rare or sparse signals in large, complex environments. Garnett's work on active search for sparse signals — cited across multiple publications — demonstrates his sustained focus on enabling aerial robots and other autonomous platforms to intelligently localize threats, detect gas leaks, or respond to emergencies by making smarter, aggregated measurements rather than exhaustive sampling. His contributions to robot grasping, reflected in his most-cited work with 19 citations, show an early and enduring interest in task-aware robotics, using graph kernels to help robots make contextually informed grasping decisions based on object properties and task constraints. His investigations into cooperative set function optimization without communication further reveal a sophisticated interest in multi-agent systems and decentralized coordination. Across these contributions, Garnett establishes himself as a researcher committed to designing intelligent systems that act efficiently and purposefully in real-world, resource-constrained environments — a body of work with meaningful implications for robotics, search-and-rescue applications, and autonomous decision systems.