Search papers, labs, and topics across Lattice.
1
0
2
Listwise preference-based reward learning from vision-language models can outperform traditional pairwise methods, achieving up to 86% success in complex robotic tasks.
