Koray Kavukcuoglu

Courant Institute of Mathematical Sciences, Princeton University, Google (United States), New York University

Papers

Total Citations

2,635

H-Index

About

Koray Kavukcuoglu is a leading figure in artificial intelligence, renowned for his foundational contributions to deep learning and computer vision. His research has fundamentally shaped how machines perceive and understand the world, with a particular focus on developing robust visual representations for autonomous systems. Kavukcuoglu’s seminal work, "Convolutional Networks and Applications in Vision," has garnered over 2,100 citations, establishing the core principles for building invariant feature representations crucial for tasks from visual perception to language understanding. He has also pioneered self-supervised learning for long-range vision in autonomous off-road driving, enabling robots to accurately classify complex terrain at distances far beyond traditional stereo systems. His innovations extend to hardware, where he developed large-scale FPGA-based convolutional networks for high-speed, low-cost embedded vision systems. Kavukcuoglu’s work on Video Pixel Networks further advanced probabilistic video modeling, capturing complex spatiotemporal dependencies in raw pixel data. A key architect of modern AI, his multi-range vision strategies for autonomous navigation have directly addressed the critical challenge of system latency in real-world robotics, cementing his legacy as a transformative force in the field.