NeurIPS 2026 Call for Papers: Robot Learning with World Models
shaohua0116 · x · 2026-08-04
NeurIPS 2026 has announced the Robot Learning with World Models workshop, exploring how world models can advance Physical AI and address key challenges in robot learning, reasoning, and control.
The workshop highlights that while state-of-the-art world models (e.g., Genie 3, Cosmos) have driven significant progress in robot reasoning, learning, and zero-shot visuomotor control, they remain primarily vision-focused. Fundamentally functioning as controllable video generation models with plausible physical dynamics, they still face persistent challenges in consistency, physical accuracy, and multi-modal physical interactions.
The event aims to bring together researchers to exchange ideas on controllable video models and discuss bridging the gap between simulated dynamics and real-world physical interactions. The deadline for paper and demo proposals is August 29, 2026.
More from Embodied
- Why Action Chunking Works in Robotic Control: Unveiling Implicit Ensembling — YouJiacheng · 2026-08-05
- NVIDIA Tutorial: Build NemoClaw Agents on Jetson Platform — NVIDIAAI · 2026-08-05
- Putting an AI Agent in a Robot Body: 4 Unexpected Lessons Learned — wolverinee04 · 2026-08-04
- Xynova's Robotic Hand Showcases Precise Pinching and Complex Finger Coordination — rohanpaul_ai · 2026-08-04
- Robotics expert: Not having legs is a safety feature for home robots — chris_j_paxton · 2026-08-04
- Agile Robots doubles revenue to €600M, signaling robotics industry growth — lukas_m_ziegler · 2026-08-04