Search papers, labs, and topics across Lattice.
4
0
6
FacePlex achieves real-time, synchronized speech and facial motion generation, outperforming existing models in lip-sync quality and motion fidelity.
Unlearning a speaker's voice from a TTS model doesn't guarantee they'll stay unlearned – current methods resurrect forgotten voices with each new unlearning request.
By separating known and unknown object representations into orthogonal subspaces, DEUS achieves state-of-the-art open world object detection, outperforming prior methods that struggle to learn distinct unknown object representations.
By treating 3D scene editing as goal-regressive planning rather than pure generation, Edit-As-Act achieves instruction fidelity, semantic consistency, and physical plausibility that existing methods miss.
