Galaxy's DexNDM enables robots to perform complex in-hand rotations & tool use (screwing, hammering) from any palm orientation, bridging the Sim2Real gap. What can't a dexterous hand do next?
#robots
open.substack.com/pub/robopub/...
#robots
open.substack.com/pub/robopub/...
Robots Master Hand Rotation
Galaxy's DexNDM enables robots to master tool use & in-hand rotation. Screwdrivers, hammers—now handled with human-like dexterity. See the future of robotics.
open.substack.com
November 10, 2025 at 8:52 AM
Galaxy's DexNDM enables robots to perform complex in-hand rotations & tool use (screwing, hammering) from any palm orientation, bridging the Sim2Real gap. What can't a dexterous hand do next?
#robots
open.substack.com/pub/robopub/...
#robots
open.substack.com/pub/robopub/...
We're working towards adding IRL stuff to LeRobot and lerobot sim2real right now
November 4, 2025 at 4:52 AM
We're working towards adding IRL stuff to LeRobot and lerobot sim2real right now
Akhil S Anand, Shambhuraj Sawant, Jasper Hoffmann, Dirk Reinhardt, Sebastien Gros: Closing the Sim2Real Performance Gap in RL https://arxiv.org/abs/2510.17709 https://arxiv.org/pdf/2510.17709 https://arxiv.org/html/2510.17709
October 21, 2025 at 6:36 AM
Akhil S Anand, Shambhuraj Sawant, Jasper Hoffmann, Dirk Reinhardt, Sebastien Gros: Closing the Sim2Real Performance Gap in RL https://arxiv.org/abs/2510.17709 https://arxiv.org/pdf/2510.17709 https://arxiv.org/html/2510.17709
Kuanning Wang, Yongchong Gu, Yuqian Fu, Zeyu Shangguan, Sicheng He, Xiangyang Xue, Yanwei Fu, Daniel Seita: SCOOP'D: Learning Mixed-Liquid-Solid Scooping via Sim2Real Generative Policy https://arxiv.org/abs/2510.11566 https://arxiv.org/pdf/2510.11566 https://arxiv.org/html/2510.11566
October 14, 2025 at 6:35 AM
Kuanning Wang, Yongchong Gu, Yuqian Fu, Zeyu Shangguan, Sicheng He, Xiangyang Xue, Yanwei Fu, Daniel Seita: SCOOP'D: Learning Mixed-Liquid-Solid Scooping via Sim2Real Generative Policy https://arxiv.org/abs/2510.11566 https://arxiv.org/pdf/2510.11566 https://arxiv.org/html/2510.11566
Kuanning Wang, Yongchong Gu, Yuqian Fu, Zeyu Shangguan, Sicheng He, Xiangyang Xue, Yanwei Fu, Daniel Seita
SCOOP'D: Learning Mixed-Liquid-Solid Scooping via Sim2Real Generative Policy
https://arxiv.org/abs/2510.11566
SCOOP'D: Learning Mixed-Liquid-Solid Scooping via Sim2Real Generative Policy
https://arxiv.org/abs/2510.11566
October 14, 2025 at 4:02 AM
Kuanning Wang, Yongchong Gu, Yuqian Fu, Zeyu Shangguan, Sicheng He, Xiangyang Xue, Yanwei Fu, Daniel Seita
SCOOP'D: Learning Mixed-Liquid-Solid Scooping via Sim2Real Generative Policy
https://arxiv.org/abs/2510.11566
SCOOP'D: Learning Mixed-Liquid-Solid Scooping via Sim2Real Generative Policy
https://arxiv.org/abs/2510.11566
A new hybrid pipeline combines online imitation pretraining in simulation with offline finetuning on limited real demos, raising robot success rates by 31.7% in sim‑to‑sim and 23.3% in sim‑to‑real transfer. https://getnews.me/sim-to-real-transfer-boosted-by-online-imitation-pretraining/ #sim2real
October 6, 2025 at 5:46 AM
A new hybrid pipeline combines online imitation pretraining in simulation with offline finetuning on limited real demos, raising robot success rates by 31.7% in sim‑to‑sim and 23.3% in sim‑to‑real transfer. https://getnews.me/sim-to-real-transfer-boosted-by-online-imitation-pretraining/ #sim2real
New #TMLR-Paper-with-Video:
REX: GPU-Accelerated Sim2Real Framework with Delay and Dynamics Estimation
Bas van der Heijden, Jens Kober, Robert Babuska, Laura Ferranti
https://tmlr.infinite-conf.org/paper_pages/O4CQ5AM5yP
#simulates #parallelization #simulated
REX: GPU-Accelerated Sim2Real Framework with Delay and Dynamics Estimation
Bas van der Heijden, Jens Kober, Robert Babuska, Laura Ferranti
https://tmlr.infinite-conf.org/paper_pages/O4CQ5AM5yP
#simulates #parallelization #simulated
October 5, 2025 at 4:30 PM
New #TMLR-Paper-with-Video:
REX: GPU-Accelerated Sim2Real Framework with Delay and Dynamics Estimation
Bas van der Heijden, Jens Kober, Robert Babuska, Laura Ferranti
https://tmlr.infinite-conf.org/paper_pages/O4CQ5AM5yP
#simulates #parallelization #simulated
REX: GPU-Accelerated Sim2Real Framework with Delay and Dynamics Estimation
Bas van der Heijden, Jens Kober, Robert Babuska, Laura Ferranti
https://tmlr.infinite-conf.org/paper_pages/O4CQ5AM5yP
#simulates #parallelization #simulated
PolySim trains policies across multiple simulators, achieving a 52.8% higher success rate on MuJoCo than IsaacSim-only training and enabling zero‑shot transfer to a Unitree G1 quadruped. Read more: https://getnews.me/polysim-reduces-sim-to-real-gap-for-humanoid-control/ #polysim #sim2real #robotics
October 3, 2025 at 10:09 PM
PolySim trains policies across multiple simulators, achieving a 52.8% higher success rate on MuJoCo than IsaacSim-only training and enabling zero‑shot transfer to a Unitree G1 quadruped. Read more: https://getnews.me/polysim-reduces-sim-to-real-gap-for-humanoid-control/ #polysim #sim2real #robotics
[2025-10-01] 📚 Updates in #LM&Ro
(1) Find the Fruit: Zero-Shot Sim2Real RL for Occlusion-Aware Plant Manipulation
(2) Multi-Robot Task Planning for Multi-Object Retrieval Tasks with Distributed On-Site Knowledge via Large Language Models
🔍 More at researchtrend.ai/communities/LM&Ro
(1) Find the Fruit: Zero-Shot Sim2Real RL for Occlusion-Aware Plant Manipulation
(2) Multi-Robot Task Planning for Multi-Object Retrieval Tasks with Distributed On-Site Knowledge via Large Language Models
🔍 More at researchtrend.ai/communities/LM&Ro
October 1, 2025 at 3:10 AM
SCOPE improves sim-to-real pose estimation, achieving a 31.9% gain on the 5°/5 cm metric and 100% grasp success on unseen object categories. Read more: https://getnews.me/scope-boosts-sim2real-object-pose-estimation-with-semantic-diffusion/ #scope #poseestimation #robotics
September 30, 2025 at 10:16 PM
SCOPE improves sim-to-real pose estimation, achieving a 31.9% gain on the 5°/5 cm metric and 100% grasp success on unseen object categories. Read more: https://getnews.me/scope-boosts-sim2real-object-pose-estimation-with-semantic-diffusion/ #scope #poseestimation #robotics
Peter H\"onig, Stefan Thalhammer, Jean-Baptiste Weibel, Matthias Hirschmanner, Markus Vincze
SCOPE: Semantic Conditioning for Sim2Real Category-Level Object Pose Estimation in Robotics
https://arxiv.org/abs/2509.24572
SCOPE: Semantic Conditioning for Sim2Real Category-Level Object Pose Estimation in Robotics
https://arxiv.org/abs/2509.24572
September 30, 2025 at 11:59 AM
Peter H\"onig, Stefan Thalhammer, Jean-Baptiste Weibel, Matthias Hirschmanner, Markus Vincze
SCOPE: Semantic Conditioning for Sim2Real Category-Level Object Pose Estimation in Robotics
https://arxiv.org/abs/2509.24572
SCOPE: Semantic Conditioning for Sim2Real Category-Level Object Pose Estimation in Robotics
https://arxiv.org/abs/2509.24572
Peter H\"onig, Stefan Thalhammer, Jean-Baptiste Weibel, Matthias Hirschmanner, Markus Vincze: SCOPE: Semantic Conditioning for Sim2Real Category-Level Object Pose Estimation in Robotics https://arxiv.org/abs/2509.24572 https://arxiv.org/pdf/2509.24572 https://arxiv.org/html/2509.24572
September 30, 2025 at 6:35 AM
Peter H\"onig, Stefan Thalhammer, Jean-Baptiste Weibel, Matthias Hirschmanner, Markus Vincze: SCOPE: Semantic Conditioning for Sim2Real Category-Level Object Pose Estimation in Robotics https://arxiv.org/abs/2509.24572 https://arxiv.org/pdf/2509.24572 https://arxiv.org/html/2509.24572
I’d planned on bringing it to life with my robots and my newly acquired dual carriage linear rail….it showed up to the studio on Monday…trying to go #Sim2Real
September 27, 2025 at 9:09 PM
I’d planned on bringing it to life with my robots and my newly acquired dual carriage linear rail….it showed up to the studio on Monday…trying to go #Sim2Real
Next Saturday, 𝗔𝗻𝘁𝗼𝗶𝗻𝗲 𝗣𝗶𝗿𝗿𝗼𝗻𝗲 will present Pollen Robotics & Hugging Face's open-source robots, including Reachy Mini, the SO-100 arm, the Amazing Hand and the Open Duck Mini. He will discuss the sim2real challenges of making the Open Duck Mini walk, and how […]
[Original post on fosstodon.org]
[Original post on fosstodon.org]
September 21, 2025 at 12:23 PM
Next Saturday, 𝗔𝗻𝘁𝗼𝗶𝗻𝗲 𝗣𝗶𝗿𝗿𝗼𝗻𝗲 will present Pollen Robotics & Hugging Face's open-source robots, including Reachy Mini, the SO-100 arm, the Amazing Hand and the Open Duck Mini. He will discuss the sim2real challenges of making the Open Duck Mini walk, and how […]
[Original post on fosstodon.org]
[Original post on fosstodon.org]
2509.11481
人間は、新しい車の運転のように、目に見えない新しい状況に適応する際、驚くほどデータ効率が高い。対照的に、強化学習（RL）を使って訓練されたニューラルネットワークポリシーのような最新のロボット制御システムは、単一の環境に高度に特化している。このようなオーバーフィッティングのため、シミュレー...
人間は、新しい車の運転のように、目に見えない新しい状況に適応する際、驚くほどデータ効率が高い。対照的に、強化学習（RL）を使って訓練されたニューラルネットワークポリシーのような最新のロボット制御システムは、単一の環境に高度に特化している。このようなオーバーフィッティングのため、シミュレー...
September 18, 2025 at 12:07 AM
2509.11481
人間は、新しい車の運転のように、目に見えない新しい状況に適応する際、驚くほどデータ効率が高い。対照的に、強化学習（RL）を使って訓練されたニューラルネットワークポリシーのような最新のロボット制御システムは、単一の環境に高度に特化している。このようなオーバーフィッティングのため、シミュレー...
人間は、新しい車の運転のように、目に見えない新しい状況に適応する際、驚くほどデータ効率が高い。対照的に、強化学習（RL）を使って訓練されたニューラルネットワークポリシーのような最新のロボット制御システムは、単一の環境に高度に特化している。このようなオーバーフィッティングのため、シミュレー...
[29/30] 95 Likes, 9 Comments, 1 Posts
2509.11481, cs․RO | cs․AI | cs․LG, 15 Sep 2025
🆕RAPTOR: A Foundation Policy for Quadrotor Control
Jonas Eschmann, Dario Albani, Giuseppe Loianno
2509.11481, cs․RO | cs․AI | cs․LG, 15 Sep 2025
🆕RAPTOR: A Foundation Policy for Quadrotor Control
Jonas Eschmann, Dario Albani, Giuseppe Loianno
September 18, 2025 at 12:07 AM
[29/30] 95 Likes, 9 Comments, 1 Posts
2509.11481, cs․RO | cs․AI | cs․LG, 15 Sep 2025
🆕RAPTOR: A Foundation Policy for Quadrotor Control
Jonas Eschmann, Dario Albani, Giuseppe Loianno
2509.11481, cs․RO | cs․AI | cs․LG, 15 Sep 2025
🆕RAPTOR: A Foundation Policy for Quadrotor Control
Jonas Eschmann, Dario Albani, Giuseppe Loianno
This research offers a framework to evaluate vision encoder pre-training, showing how manipulation-pretrained models enhance Sim2Real transfer for robotic policies. It underscores feature extraction and domain invariance to improve robotic learning in real settings. https://arxiv.org/abs/2501.16389
Bridging the Sim2Real Gap: Vision Encoder Pre-Training for Visuomotor Policy Transfer
ArXiv link for Bridging the Sim2Real Gap: Vision Encoder Pre-Training for Visuomotor Policy Transfer
arxiv.org
September 10, 2025 at 7:20 AM
This research offers a framework to evaluate vision encoder pre-training, showing how manipulation-pretrained models enhance Sim2Real transfer for robotic policies. It underscores feature extraction and domain invariance to improve robotic learning in real settings. https://arxiv.org/abs/2501.16389
Was reading some sim2real code the other day and getting LLM to make a little tutorial on PPO and it starts with "PPO is an actor-critic method" and I'm like hold tf up no tf it is not. And it's like "yes it is" and I'm like "I read the original paper by Shulman it's a policy not a value method"
September 7, 2025 at 10:46 AM
Was reading some sim2real code the other day and getting LLM to make a little tutorial on PPO and it starts with "PPO is an actor-critic method" and I'm like hold tf up no tf it is not. And it's like "yes it is" and I'm like "I read the original paper by Shulman it's a policy not a value method"
Muhammad Shahbaz, Shaurya Agarwal
PercepTwin: Modeling High-Fidelity Digital Twins for Sim2Real LiDAR-based Perception for Intelligent Transportation Systems
https://arxiv.org/abs/2509.02903
PercepTwin: Modeling High-Fidelity Digital Twins for Sim2Real LiDAR-based Perception for Intelligent Transportation Systems
https://arxiv.org/abs/2509.02903
September 4, 2025 at 7:46 AM
Muhammad Shahbaz, Shaurya Agarwal
PercepTwin: Modeling High-Fidelity Digital Twins for Sim2Real LiDAR-based Perception for Intelligent Transportation Systems
https://arxiv.org/abs/2509.02903
PercepTwin: Modeling High-Fidelity Digital Twins for Sim2Real LiDAR-based Perception for Intelligent Transportation Systems
https://arxiv.org/abs/2509.02903
Muhammad Shahbaz, Shaurya Agarwal
High-Fidelity Digital Twins for Bridging the Sim2Real Gap in LiDAR-Based ITS Perception
https://arxiv.org/abs/2509.02904
High-Fidelity Digital Twins for Bridging the Sim2Real Gap in LiDAR-Based ITS Perception
https://arxiv.org/abs/2509.02904
September 4, 2025 at 7:45 AM
Muhammad Shahbaz, Shaurya Agarwal
High-Fidelity Digital Twins for Bridging the Sim2Real Gap in LiDAR-Based ITS Perception
https://arxiv.org/abs/2509.02904
High-Fidelity Digital Twins for Bridging the Sim2Real Gap in LiDAR-Based ITS Perception
https://arxiv.org/abs/2509.02904
Muhammad Shahbaz, Shaurya Agarwal: High-Fidelity Digital Twins for Bridging the Sim2Real Gap in LiDAR-Based ITS Perception https://arxiv.org/abs/2509.02904 https://arxiv.org/pdf/2509.02904 https://arxiv.org/html/2509.02904
September 4, 2025 at 6:30 AM
Muhammad Shahbaz, Shaurya Agarwal: High-Fidelity Digital Twins for Bridging the Sim2Real Gap in LiDAR-Based ITS Perception https://arxiv.org/abs/2509.02904 https://arxiv.org/pdf/2509.02904 https://arxiv.org/html/2509.02904
Muhammad Shahbaz, Shaurya Agarwal: PercepTwin: Modeling High-Fidelity Digital Twins for Sim2Real LiDAR-based Perception for Intelligent Transportation Systems https://arxiv.org/abs/2509.02903 https://arxiv.org/pdf/2509.02903 https://arxiv.org/html/2509.02903
September 4, 2025 at 6:30 AM
Muhammad Shahbaz, Shaurya Agarwal: PercepTwin: Modeling High-Fidelity Digital Twins for Sim2Real LiDAR-based Perception for Intelligent Transportation Systems https://arxiv.org/abs/2509.02903 https://arxiv.org/pdf/2509.02903 https://arxiv.org/html/2509.02903
[2025-09-04] 📚 Updates in #3DV
(1) Refinement of Monocular Depth Maps via Multi-View Differentiable Rendering
(2) PercepTwin: Modeling High-Fidelity Digital Twins for Sim2Real LiDAR-based Perception for Intelligent Transportation Systems
🔍 More at researchtrend.ai/communities/3DV
(1) Refinement of Monocular Depth Maps via Multi-View Differentiable Rendering
(2) PercepTwin: Modeling High-Fidelity Digital Twins for Sim2Real LiDAR-based Perception for Intelligent Transportation Systems
🔍 More at researchtrend.ai/communities/3DV
September 4, 2025 at 3:07 AM
It really is, and the underlying tech making these things possible has so many uses. "Style transfer" image models are super useful at compensating for small real datasets and sim2real
September 1, 2025 at 8:42 PM
It really is, and the underlying tech making these things possible has so many uses. "Style transfer" image models are super useful at compensating for small real datasets and sim2real