Gabriele Rosi
Papers
1
Total Citations
6
H-Index
1
About
Gabriele Rosi is a rising researcher at the intersection of computer vision, robotics, and cognitive science, with a focused interest in visual affordance reasoning. His most-cited work, "What does CLIP know about peeling a banana?" (2024, 6 citations), investigates how large vision-language models can be leveraged to segment object parts based on the actions they afford—a critical capability for enabling robots to interact intelligently with everyday objects. This research bridges the gap between human-like understanding of tool use and robotic manipulation, addressing a fundamental challenge in embodied AI. By probing what models like CLIP inherently "know" about functional object properties, Rosi’s work contributes to more intuitive human-robot interaction and autonomous task execution. Though early in his career, his focus on grounding abstract visual knowledge in actionable, part-based segmentation marks a promising direction for the field. His research holds particular relevance for students and engineers working on robotic perception, scene understanding, and the integration of foundation models into physical systems.
Research Focus
Key Achievements
Top Papers
- 1What does CLIP know about peeling a banana?6 citations · 2024