Search papers, labs, and topics across Lattice.
5
37
5
11
Scaling visuomotor context to 8K timesteps enables robots to master complex tasks and adapt in real-time, outperforming previous models by a staggering margin.
VLN agents can navigate more accurately in zero-shot settings by "looking forward, now, and backward," mimicking human navigational strategies.
VLA-Thinker lets robots actively "look again" at their environment during long tasks, leading to a huge performance boost in complex manipulation.
Forget painstakingly engineering robot behaviors: DreamZero learns directly from video of other robots or even humans, adapting to new tasks and bodies with just minutes of data.
Forget synthetic data that looks like it came from a PS2 game: NVIDIA's new Cosmos-Predict2.5 generates high-fidelity videos for training embodied AI, opening the door to more realistic and reliable simulations.
