معرفی
Dr. Danfei Xu is an Assistant Professor at the School of Interactive Computing at Georgia Tech and a part-time Research Scientist at NVIDIA AI. He holds a B.S. from Columbia University (2015) and a Ph.D. from Stanford University (2021), advised by Fei-Fei Li and Silvio Savarese. His research focuses on enabling robots to perform physical tasks in human environments through advancements in robotics, machine learning, and computer vision.
Education:
- B.S. in Computer Science, Columbia University (2015)
- Ph.D. in Computer Science, Stanford University (2021)
Research Interests:
- Robot Learning from Human Data (e.g., egocentric videos)
- Long-horizon reasoning via generative models
- Robot learning systems (hardware/software integration)
- Imitation learning, reinforcement learning, and visuomotor control
Recent Articles Trends:
His publications emphasize generative models for skill chaining, uncertainty-driven mapping, and human-in-the-loop learning. Key themes include leveraging diffusion models for complex manipulation tasks and developing open-source robotic systems.
Awards and Grants:
- Meta Grant (2024) for robot learning from human data
- NSF Grant (2024) for generative models in robotics
- Autodesk Grant (2023) for high-precision manipulation
- IEEE RA-L Best Paper Honorable Mention (2024)
- ICRA 2024 Best Paper Award (Open-X-Embodiment)
Advising and Teaching:
- Georgia Tech courses: Deep Learning (CS 7643/4644), Deep Learning for Robotics (CS 8803 DLM)
- Ph.D. recruitment in Robotics/ML/AI
- Lab: Robot Learning and Reasoning Lab (RL2)
Labs and Teams:
- RL2 Lab develops adaptable robotic 'brains' for home, healthcare, and search/rescue
- Focus areas: End-to-end learning, generative models, and open-source systems
