Search papers, labs, and topics across Lattice.
3
0
6
0
Implicit reward models can now more accurately pinpoint correct reasoning steps, thanks to a novel prefix-value learning approach that closes the train-inference gap.
Music-grounded video editing can now produce significantly more coherent timelines thanks to a novel global-local coordination mechanism that resolves cross-segment conflicts.
Zero-shot 3D visual grounding can be achieved by decoupling spatial semantics (resolved by 2D VLMs) from 3D structure instantiation (handled by deterministic multi-view geometry), outperforming even fully supervised methods.
