Matthew Schmittle

Papers

Total Citations

H-Index

About

Matthew Schmittle is a leading researcher at the intersection of robotics and machine learning, specializing in scalable human-robot interaction and imitation learning. His work fundamentally addresses a critical bottleneck in robotics: how to enable robots to learn complex tasks efficiently from non-expert human teachers. Schmittle’s major contribution lies in developing frameworks that move beyond static, pre-recorded demonstrations. His most-cited work, "Learning from Interventions" (42 citations), pioneers a paradigm where a robot learns from both explicit and implicit human feedback during live interaction, making the teaching process more intuitive and seamless. He further advanced this with "Expert Intervention Learning" (27 citations), which formalizes how robots can leverage corrective interventions to improve policy learning. His research directly tackles the "optimal demonstration problem"—the difficulty humans face in providing perfect kinesthetic demonstrations—by allowing robots to learn from imperfect, corrective feedback. Through these innovations, Schmittle is paving the way for robots that can be taught by anyone, anywhere, dramatically reducing the barrier to deploying adaptable robotic systems in real-world environments.