Search papers, labs, and topics across Lattice.
S1 alone, (ii) S1+S, Shanghai Innovation Institute
2
0
5
4
LLMs can now plan complex, long-horizon household tasks from ambiguous instructions with significantly improved performance, thanks to a novel reinforcement learning approach that corrects intermediate reasoning steps.
Achieve one-shot skill transfer in imitation learning by disentangling intent from execution, allowing robots to adapt to new environments simply by injecting the "Intent" token from a demonstration.
