Joy Hsu
Papers
2
Total Citations
7
H-Index
2
About
Joy Hsu is a rising star in the field of robotics and artificial intelligence, whose research sits at the exciting intersection of robotic manipulation, visual reasoning, and foundation models. Her work tackles a fundamental challenge: enabling robots to not only perform complex physical tasks but also to semantically reason about *when* and *how* to apply those skills. In her highly-cited 2023 paper, "Programmatically Grounded, Compositionally Generalizable Robotic Manipulation," she pioneered a method that integrates large-scale pretrained vision-language models into robotic systems, allowing for rich, compositionally generalizable manipulation. This work, already garnering significant attention, demonstrates how robots can break down novel tasks into reusable skills. Complementing this, her paper "What's Left? Concept Grounding with Logic-Enhanced Foundation Models" pushes the boundaries of visual reasoning by using large language models to compose programs for vision-language models, moving beyond simple 2D images toward more robust, logic-grounded understanding. With over 7 citations across her early publications, Hsu is rapidly establishing herself as a key innovator in creating more intelligent, adaptable, and semantically-aware robots.
Research Focus
Key Achievements
Top Papers
- 1
- 2What's Left? Concept Grounding with Logic-Enhanced Foundation Models3 citations · 2023