Search papers, labs, and topics across Lattice.
5
0
7
14
Language instructions now directly control real-time, high-fidelity 360掳 video generation, enabling fluid interactive exploration from a single panorama.
Current generative models struggle with temporal drift and responsiveness in streaming audio-video generation, as revealed by the new StreamAV-Bench benchmark.
Explicitly disentangling world dynamics from view rendering allows MAS to achieve unprecedented scalability and consistency in multiplayer simulations.
Modern video generation models can inherently estimate dynamic lighting, producing HDR environment maps that are both temporally coherent and physically plausible.
Finally, video generation achieves granular control over scene lighting, layout, and camera movement, unlocking new creative possibilities.
