Nico Hauff

Intelligent Systems Research (United States), University of Freiburg

Papers

Total Citations

H-Index

About

Nico Hauff is a robotics researcher whose work lies at the intersection of computer vision and robotic manipulation, with a particular focus on enabling robots to interact with unknown objects in unstructured environments. His major contributions center on developing self-supervised learning methods that allow robots to learn from physical interaction rather than relying on large, hand-labeled datasets. In his highly cited 2019 paper on self-supervised transfer learning for instance segmentation, Hauff demonstrated how robots can use physical interaction—such as pushing or grasping—to generate their own training data, achieving robust segmentation performance without manual annotation. This work has garnered 24 citations and addresses a critical bottleneck in robotic perception. Hauff further advanced the field with his work on learning to singulate objects using a push proposal network, which enables robots to separate individual items from cluttered piles—a fundamental skill for tasks like sorting and assembly. By combining deep learning with interactive perception, Hauff’s research offers scalable solutions for real-world robotic systems, making him a notable contributor to the growing field of robot learning from interaction.