Jeremy Stober

Papers

Total Citations

H-Index

About

Jeremy Stober’s research sits at the intersection of robotics, cognitive science, and developmental learning, with a focus on how autonomous systems can discover structure from raw sensorimotor experience. His work explores a fundamental challenge: how a robot—like a human infant—can make sense of a “blooming, buzzing confusion” of sensory data without pre-programmed models. In his most cited paper, “Learning geometry from sensorimotor experience” (2011, 13 citations), Stober demonstrates how a robot can infer geometric properties of its body and environment purely through interaction, bypassing the need for manual calibration. His earlier work on “Sensor Map Discovery for Developing Robots” (2009, 12 citations) tackles the tedious process of calibrating complex sensor arrays—cameras, lasers, and sonars—by enabling robots to autonomously learn sensor geometry and behavior. In “Learning the Sensorimotor Structure of the Foveated Retina” (2009, 7 citations), Stober draws inspiration from human vision, showing how foveated sensing and saccadic eye movements can jointly teach a system to learn receptive field structures and attention policies. This biologically-inspired approach not only advances robotic autonomy but also offers insights into developmental cognition. Stober’s contributions are foundational for building robots that learn like living organisms, reducing the burden of manual engineering in favor of self-discovery.