Skip to content

Learning andPerceptionResearch

Advancing Embodied and Scientific Intelligence

Embodied and Scientific Intelligence

NVIDIA’s Learning and Perception Research (LPR) group, led by Jan Kautz, advances embodied intelligence for humanoid robots, digital agents, and AI scientists. Our teams develop robot policies, efficient foundation models, and robust perception and reasoning systems. We publish our research and openly release code and models for the broader community.

LPR’s Teams

Recent Publications

View all 426

DiLaDiff: Distilled Latent-augmented Diffusion for Language Modeling

Jean-Marie Lemercier, Tomas Geffner, Morteza Mardani, Karsten Kreis, Arash Vahdat, Ante Jukić

Advances in Neural Information Processing Systems (NeurIPS) · 2026

Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention

Ali Hatamizadeh, Yejin Choi, Jan Kautz

Advances in Neural Information Processing Systems (NeurIPS) · 2026

PointZero: 3D Point Track Completion for Learning Transferable 3D Dynamics

Bardienus Pieter Duisterhof, Kaifeng Zhang, Adam Hung, Bowen Wen, Stan Birchfield, Yunzhu Li, Deva Ramanan, Jeffrey Ichnowski

Advances in Neural Information Processing Systems (NeurIPS) · 2026

Test-time Scaling of Diffusions with Flow Maps

Amirmojtaba Sabour, Michael Samuel Albergo, Carles Domingo-Enrich, Nicholas Matthew Boffi, Sanja Fidler, Karsten Kreis, Eric Vanden-Eijnden

Advances in Neural Information Processing Systems (NeurIPS) · 2026

Addressable Memory for Video World Models

Xindi Wu, Sven Elflein, James Lucas, Olga Russakovsky, Laura Leal-Taixé, Despoina Paschalidou, Jonathan Lorraine, Aljoša Ošep

Advances in Neural Information Processing Systems (NeurIPS) · 2026

Nemotron-Labs-Diffusion-Image: Advancing Masked Discrete Diffusion for High-Resolution Image Synthesis

Shufan Li, Greg Heinrich, Hanrong Ye, Yonggan Fu, Aditya Grover, Jan Kautz, Pavlo Molchanov

Advances in Neural Information Processing Systems (NeurIPS) · 2026