Search papers, labs, and topics across Lattice.
University of Maryland
3
0
8
Just 30 minutes of human egocentric video can train robots to perform complex manipulation tasks with superhuman dexterity.
Counterintuitively, throwing more user prompts at prompt optimization can backfire, especially on diverse datasets, but a clever filtering strategy can unlock surprisingly strong generalization from just a handful of examples.
Forget fancy architectures: simply switching from Adam to SGD can boost Vision Transformer performance by +25 to +38 F1 points in post-tornado damage recognition.
