#RobotLearning
LightNews LightNews
robotics.biz
NVIDIA just opened its newest robot brain at #CoRL2026.

GR00T 2 is open, trained on 100,000 hours of human video. With one demo per task it hit 71% across 50 tasks; with five, 95%. Foxconn and Siemens are already testing it.

Source: NVIDIA Blog
Video: NVIDIA

#Robotics #RobotLearning
November 9, 2026 at 6:04 PM
susanfourtane.bsky.social
I’m looking forward to meeting with #robotics #startups, #founders, and #investors this month during the #Slush week in #Helsinki. Exciting robotics demos this year! #AI #industrial #manufacturing #robotlearning #automation #sensors #tech #machinevision

susanfourtane.substack.com/p/can-robots...
Can robots really learn a new task after seeing it just once?
A London-based robotics startup develops robots capable of learning new physical tasks from a single demo. This could accelerate the arrival of adaptable, intelligent machines in factories before 2030
susanfourtane.substack.com
November 1, 2026 at 4:31 PM
utiasstars.bsky.social
6/ Come say hi if you're at CoRL this year. See you in Austin 🤠

🌐 papers.starslab.ca/diffusion-ik...
📄 arxiv.org/abs/2606.26408
💻 github.com/utiasSTARS/C...

#RobotLearning #DiffusionModels #UofT
October 31, 2026 at 7:18 PM
robotics.biz
Show Machenta a task once. The robot does it.

The Imperial spinout raised $40M, a record seed for a UK robotics startup. Trained only in simulation, its model learns from that one demo. No real-world data.

Source: Imperial College London
Video: Machenta

#Robotics #RobotLearning #PhysicalAI
November 1, 2026 at 1:22 AM
robotics.biz
Bear Robotics is moving from carrying goods to handling them. #Humanoid

KINO-2 teaches its KR1 humanoid manipulation from demos, cutting GR00T inference from 500ms to 60ms. Bear runs 16,000+ robots at 5,000+ sites.

Source: Business Wire
Image: Bear Robotics

#Robotics #PhysicalAI #RobotLearning
October 27, 2026 at 9:45 PM
trossenrobotics.bsky.social
99% accurate robot, 36% chance it finishes the job. The math behind why teleop is permanent. #robotics #AI #teleop #robotlearning
October 24, 2026 at 3:02 PM
robotics.biz
NVIDIA just open-sourced Flora, its GEAR lab model suite that turns raw video into robot training data

It labels 200K hours of footage a day in house at a claimed 125x lower cost. The weights are now public.

Source: NVIDIA GEAR
#Robotics #RobotLearning #EmbodiedAI
October 20, 2026 at 9:19 PM
robotics.biz
$150M for teaching robots to feel. Eka Robotics closed a Series A at a $1.45B valuation, led by a16z and F-Prime.

Its force-sensing gripper picks 1,500 objects an hour, 2.5x human speed, without crushing fragile ones.

Source: TechCrunch
Video: Eka Robotics

#Robotics #PhysicalAI #RobotLearning
October 20, 2026 at 5:25 PM
robotics.biz
One policy, twenty assembly tasks. AlphaNIST, from NVIDIA's Isaac Lab, trained a single RL policy across the full NIST assembly board: 98.2% success vs 81.4% for specialist policies. It even chains 10 tasks in a row on a single reward. Source: AlphaNIST. Video: AlphaNIST #Robotics #RobotLearning
October 15, 2026 at 11:54 PM
robotcurrent.bsky.social
HumanX is a new framework compiling monocular human video into generalizable interaction skills for humanoid robots, combining an anchor-centric data pipeline with an interaction-prioritized learning…

#HumanX #humanoidRobots #robotLearning #imitationLearning
https://arxiv.org/abs/2602.02473
October 12, 2026 at 4:01 AM
artestomellivoura.bsky.social
0HFLOW: OPEN-SOURCE DATA PIPELINES POWERING THE FUTURE OF ROBOTICS & PHYSICAL AI 🤖⚙️

#HFlow #HebbianRobotics #PhysicalAI #Robotics #ArtificialIntelligence #OpenSource #MachineLearning #RoboticsEngineering #AITraining #DataEngineering #DataPipelines #RobotLearning #Automation #Python #FutureOfAI
October 9, 2026 at 6:40 PM
tech-journal.bsky.social
Welcome to read the cover paper from Prof. Huijun Gao's team from Harbin Institute of Technology!
Title: #Embodied #AI for #aerial #manipulation: a survey
Keywords: #embodiedintelligence #AI #aerialmanipulation #robotlearning

Get the full text here: technicapub.com/technology/a...
October 9, 2026 at 12:57 PM
mlscientist.bsky.social
KTH Royal Institute of Technology is offering a two-year postdoctoral position in robot learning for manipulation using Vision-Language-Action models.

mlscientist.com/postdoc-robo...

#robotlearning
October 7, 2026 at 6:08 PM
promptfoundry.bsky.social
A new episodic, derivative-free latent policy improvement approach boosts pretrained diffusion or flow matching robot policies by adjusting initial noise sampling, working across MLP and VLA models…

#RobotLearning #DiffusionModels #PolicyImprovement #GenerativeAI
https://arxiv.org/abs/2603.15757
October 6, 2026 at 4:01 PM
trossenrobotics.bsky.social
Demos recorded. Model training. Laptop closed.

Whatever the results look like, that's for Monday.

Happy Friday from all of us at Trossen. What's training in your lab this weekend?

#FridayFeeling #Robotics #PhysicalAI #RobotLearning
October 2, 2026 at 10:00 PM
trossenrobotics.bsky.social
"Surreal." 🦾

Move your hand. The robot glides right along with you. Smooth, natural, no lag. That's Glide and Cockpit. Link in bio to go to the website and checkout our new product lineup.

#PhysicalAI #Teleoperation #Robotics #RobotArms #RobotLearning #FirstReactions
September 30, 2026 at 9:52 PM
vicky-grok.bsky.social
#Robotics #AI #FacebookResearch #ReinforcementLearning #Simulation #OpenSource #RobotLearning #MachineLearning
September 16, 2026 at 1:39 PM
markusknr.wsocial.eu
Next week I will head to IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) #IROS2026 in Pittsburgh. I'll be there for the full week, 27 Sep – 1 Oct, workshops included, and would love to connect. Here's where you'll definitely find me.

#Robotics #RobotLearning #HRI #VLM
September 15, 2026 at 7:11 PM
markusknauer.bsky.social
Next week I will head to IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) #IROS2026 in Pittsburgh. I'll be there for the full week, 27 Sep – 1 Oct, workshops included, and would love to connect. Here's where you'll definitely find me.

#Robotics #RobotLearning #HRI #VLM
September 15, 2026 at 7:11 PM
manavpod.bsky.social
GEN-1.5 learns new tasks from a single 12-second demo.

Built by Generalist AI, it skips weeks of programming entirely. 

Show it once, and it immediately attempts the task with zero retraining, achieving 59% success rate instantly.

#GEN15 #GeneralistAI #RobotLearning
August 26, 2026 at 11:03 AM
aidailypost.com
New research shows a Generalist AI (GEN‑1.5) can teach robots fresh tasks with 83% success in just 10 steps—thanks to clever physical prompts and zero‑shot learning. Curious how context windows make it happen? Dive in! #GEN15 #GeneralistAI #RobotLearning

🔗 aidailypost.com/news/ai-teac...
August 20, 2026 at 12:50 PM
detoxima2025.bsky.social
로봇 학습 데이터로 공짜 청소 받는 법 3가지

https://bit.ly/4vjpo5T

#로봇학습 #인공지능 #데이터수집 #가사로봇 #RobotLearning #ArtificialIntelligence #DataCollection
로봇 학습 데이터로 공짜 청소 받는 법 3가지 - 천의무봉
최근 내 집을 무료로 청소해 주는 대신 일상적인 가사 노동 모습을 촬영해 가겠다는 독특한 서비스가 등장했습니다. 얼핏 솔깃한 제안 같지만 여기에는 테크 기업들의 치밀한 계산이 깔려 있습니다. 바로 인공지능이 인간 세상을 배우기 위해 필요한 로봇 학습 데이터 확보 경쟁 때문입니다.
blog.hangadac.com
May 29, 2026 at 7:53 PM
nguyensmai.bsky.social
Proud of Mehdi's successful PhD presentation 🎓 of his work combining #ReinforcementLearning and #FormalMethods for robots to learn online Automatic Symbolic Goal Abstraction via Reachability Analysis in Hierarchical Reinforcement Learning
📽️ youtu.be/vOjRKN8MFAo
#EmergentSymbols #RobotLearning
Automatic Symbolic Goal Abstraction via ReachabilityAnalysis in Hierarchical Reinforcement Learning
YouTube video by Nguyen Sao Mai
youtu.be
September 3, 2025 at 2:53 PM
nguyensmai.bsky.social
I finally uploaded my presentation "The #IntrinsicMotivation of #ReinforcementLearning and #ImitationLearning for sequential tasks" which overviews my past 10 years of research
📽️ youtu.be/tPndN5dntoQ
#ActiveImitationLearning #CompositionalLearning #ActivityRecognition #ADL #RobotLearning #RobotCoach
The intrinsic motivation of reinforcement and imitation learning for sequential tasks
YouTube video by Nguyen Sao Mai
youtu.be
August 24, 2025 at 2:06 PM
georgiachal.bsky.social
🚨 We're hiring a Postdoc in Robot Learning @ PEARL Lab, TU Darmstadt 🚨

Join our ERC-funded project SIREN (Structured Interactive Perception and Learning for Holistic Robotic Embodied Intelligence).

#robotics #AI #postdoc #ERCgrant #RobotLearning #EmbodiedAI #AIRobots
#MachineLearning#roboticsjobs
Open positions
We are seeking for an excellent Postdoc to join our group and contribute to the ERC StG SIREN. Job description available here We are seeking for enthusiastic researchers that want to work on highly…
pearl-lab.com
May 25, 2025 at 9:24 AM