MAGIC^VFM — Meta-Learning with Vision Foundation Models

MAGIC^VFM is a meta-learning framework that adapts robot ground interaction control using visual foundation models. The system enables robots to quickly adapt their locomotion strategy to different terrain types by leveraging rich visual representations from large pre-trained models.

Publication: Lupu, E. S.*, Xie, F.*, Preiss, J., Anderson, M., Alindogan, J., & Chung, S.-J. (2024). MAGIC^VFM: Meta-learning adaptation for ground interaction control with visual foundation models. IEEE Transactions on Robotics.