Sanja Fidler

GRAIL achieves an impressive 84% success rate in real-world object pick-up tasks using only synthetic data, revolutionizing humanoid robot training.

A real-time generative world model can synthesize complex driving scenarios that traditional simulators struggle to capture, enabling safer and more effective evaluation of autonomous vehicle policies.

Cosmos 3 sets a new benchmark for omnimodal models, outperforming existing state-of-the-art in both Text-to-Image and Image-to-Video tasks.

Forget fixed agent slots and quadratic attention: Gamma-World uses simplex embeddings and sparse hubs to generate interactive multi-agent environments with better fidelity and control, even generalizing from 2 to 4 players without retraining.

Text-to-image models can now generate megapixel images 6x faster and with better quality by replacing traditional decoders with a pixel diffusion-based upsampler.

Transforming sparse driving log observations into complete 3D assets could revolutionize simulation fidelity in autonomous vehicle development.

Forget generating static 3D scenes – Lyra 2.0 lets you create entire explorable 3D worlds by cleverly routing information from past frames and training the model to correct its own mistakes.

Finally, a video generation model lets you puppeteer objects and their reactions independently, all while freely moving the camera.

Kimodo leaps ahead in controllable human motion generation by training a diffusion model on a massive 700-hour mocap dataset, enabling unprecedented control fidelity via text and diverse kinematic constraints.

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.