Michael Pearce

University of Warwick, Georgia Institute of Technology

Papers

Total Citations

276

H-Index

About

Michael Pearce is a versatile researcher whose work spans two distinct yet complementary domains: machine learning optimization and autonomous robotics. His most influential contribution, "Scalable Global Optimization via Local Bayesian Optimization" (2019), has garnered 144 citations and addresses one of the field's most pressing challenges — extending Bayesian optimization to high-dimensional problems with thousands of observations, where traditional approaches often falter. This work has significantly advanced the practical applicability of sample-efficient optimization for expensive black-box functions across real-world engineering and scientific applications. Earlier in his career, Pearce made pioneering contributions to evolutionary robotics, with his 1994 paper on applying genetic algorithms to autonomous robot navigation accumulating 110 citations — a remarkable impact for work of that era. By evolving reactive control systems across diverse environments to create adaptable "ecological niches," he helped lay foundational principles for intelligent robotic behavior. A follow-up publication in 2005 further consolidated these ideas for broader audiences. Together, Pearce's body of work reflects a sustained commitment to making intelligent systems more adaptive and efficient, bridging classical evolutionary computation with modern probabilistic optimization — a trajectory that continues to influence both robotics and machine learning researchers today.