SIM1 (9/9)
💖 Acknowledgements.
This work is built upon and inspired by the collective progress of the robot learning and embodied intelligence community. We sincerely thank the following teams and projects for their foundational insights:
· NVIDIA Robotics & NVIDIA Omniverse team (
@NVIDIARobotics,
@nvidiaomniverse): for advancing large-scale simulation infrastructure and robotics-oriented digital twin ecosystems.
· Style3D (
@style3dofficial): for valuable insights on digital garment modeling and simulation pipelines, and for the exciting ongoing collaboration.
· Lightwheel AI (
@LightwheelAI): for inspiring perspectives on scalable embodied intelligence and data-driven robot learning systems.
· GarmentLab (
@warshallrho): for pioneering work on deformable garment manipulation, which strongly influenced our problem formulation.
· OpenDriveLab (Kai0 series) (
@OpenDriveLab,
@jiazhi_yang2024,
@smch_1127): for early and influential explorations of real-world deformable manipulation systems and benchmarks.
· MimicGen (
@yukez, Dieter Fox): for foundational work on scalable robot data generation via imitation and compositional task synthesis, which strongly informed our data engine design.
· Chelsea Finn & Pieter Abbeel (
@chelseabfinn,
@pabbeel): for foundational advances in policy learning and advantage-based formulations that underpin modern imitation learning systems; we build upon policy families such as Pi0 / Pi0.5 as strong baselines.
· Jim Fan (
@DrJimFan) and the DreamZero team (
@jang_yoel,
@ShenyuanGao): for shaping the field’s perspective on scaling robot learning with video and generative models, and for inspiring our thinking on data scaling regimes.