Search papers, labs, and topics across Lattice.
3
0
5
DenseReward synthesizes diverse failure trajectories automatically, enabling robots to learn from a rich array of failure modes without human labeling.
By decomposing long-horizon manipulation into transport and object-centric interaction, LiLo-VLA achieves state-of-the-art zero-shot generalization and robustness, outperforming end-to-end VLA models by a large margin.
VLAs often ignore your instructions and just do what they've seen before, but a simple "counterfactual comparison" trick can fix it.
