Track trials, findings, and clinical evidence as it lands.
Search papers, labs, and topics across Lattice.
University of Southern California
3
0
5
AI agents can now generate and verify complex 3D worlds using executable code, outperforming leading models in human preference studies.
VLMs struggle with procedural 3D modeling, often producing flawed outputs due to API mismatches and geometric disconnections, but performance can be significantly boosted through iterative refinement.
VLN agents can navigate more accurately in zero-shot settings by "looking forward, now, and backward," mimicking human navigational strategies.