Andrea Vedaldi

Papers

Total Citations

197

H-Index

About

Andrea Vedaldi is a leading figure in computer vision and machine learning, whose research bridges the gap between visual perception and autonomous decision-making. His work is centered on developing algorithms that allow machines to understand and interact with complex, dynamic environments. A key contribution is his pioneering work on spatial memory for autonomous agents, exemplified by the highly influential "MapNet" (2018, 162 citations), which introduced an allocentric spatial memory system enabling agents to build and reason about world maps from egocentric sensory input—a fundamental step for long-term navigation and scene understanding. Vedaldi has also advanced self-supervised learning for 3D vision, notably in monocular depth estimation, where his methods allow accurate depth reconstruction from raw video without costly 3D ground truth, a crucial capability for robotics. As the editor of a seminal "Deep Learning for Computer Vision" volume (2017, 13 citations), he has helped shape the field’s pedagogical foundations. His more recent work on goal-conditioned visuomotor control (2021) pushes toward versatile robotic skill primitives, demonstrating his sustained impact on embodied AI. With over 160 citations on his core spatial memory work alone, Vedaldi’s research continues to define how machines perceive, remember, and act in the world.