Search papers, labs, and topics across Lattice.
1
0
3
Forget everything you thought you knew about continual learning: pretrained Vision-Language-Action models can learn new robotic skills without catastrophic forgetting, even with minimal replay.
