Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
3
0
6
9
AI agents can now generate and verify complex 3D worlds using executable code, outperforming leading models in human preference studies.
VLMs can learn to actively explore 3D worlds and acquire visual evidence for reasoning, achieving 53.2% accuracy by optimizing for final-answer correctness alone.
A 1000x larger video reasoning dataset reveals early signs of emergent generalization, offering a new foundation for training and evaluating spatiotemporal AI.