Sandeep Bajamahal @ ICML🇰🇷 retweetledi
anthropic.com/research/claud…
Claude plays robotics
A must-read!!!
English
Sandeep Bajamahal @ ICML🇰🇷
13 posts
Sandeep Bajamahal @ ICML🇰🇷
@sandeep_b24
EECS @UCBerkeley, Intern @nvidia, Undergrad Researcher @AUTOlab_Cal @berkeley_ai
Katılım Ağustos 2025
68 Takip Edilen27 Takipçiler
Excited to share our latest work focusing on agentic robotics!
Kaiyuan Eric Chen ✈️RSS🇦🇺@keplerccccc
Graph-as-Policy (GaP) is a new variant of Agentic Robotics from @NVIDIA and @UCBerkeley that builds computation graphs (like ROS) to ensure modularity, manage complexity, and facilitate interpretability. 🧵Open code and paper: graph-robots.github.io/gap/ #Robotics #Automation #AgenticAI #EmbodiedAI
English
Sandeep Bajamahal @ ICML🇰🇷 retweetledi
Introducing WARP-RM: A Warp-Augmented Relative Progress Reward Model for Data Curation 🧵
We gave our t-shirt folding robot more demonstrations and it got worse. Every extra demo ended in a successfully folded shirt. The data wasn't bad. It was noisy. The policy couldn't tell productive motion from dead time, and it imitated both equally. So which moments of a demo are actually worth copying?
🌐 Project Website: uynitsuj.github.io/warp-rm
📄 Paper: arxiv.org/abs/2606.28320
💻 Code: github.com/uynitsuj/WARP-…
📨 XDOF blog post: xdof.ai/blog/warp-rm
English
Sandeep Bajamahal @ ICML🇰🇷 retweetledi
Sandeep Bajamahal @ ICML🇰🇷 retweetledi
Sandeep Bajamahal @ ICML🇰🇷 retweetledi
Glad to share that OXE-AugE has been accepted as a spotlight paper (2.2%) at #ICML2026!
Lawrence Yunliang Chen@Lawrence_Y_Chen
Can robot data evolve with robot hardware? Robot arms and hands are rapidly evolving based on advances in materials, motors, sensors, mechanical designs, and applications. Now, demo data can also evolve: AugE-Toolkit is a new open-source, easy-to-use, scalable package for augmenting robot embodiments: it converts demo data from one robot arm/gripper to a different robot arm/gripper in a scalable way. We used AugE-Toolkit to create OXE-AugE––a large open-source dataset that augments 16 popular OXE subsets with 9 different robot embodiments to provide over 2M new trajectories, tripling the size of the original dataset. The OXE-AugE dataset is on HuggingFace and you can download and train right away! Most importantly, we find robot augmentation scales! 🌐 oxe-auge.github.io 🤗 huggingface.co/oxe-auge 🔗arxiv.org/abs/2512.13100 🧵👇 (1/9)
English
Sandeep Bajamahal @ ICML🇰🇷 retweetledi
Robotics: coding agents’ next frontier.
So how good are they?
We introduce CaP-X: an open-source framework and benchmark for coding agents, where they write code for robot perception and control, execute it on sim and real robots, observe the outcomes, and iteratively improve code reliability.
From @NVIDIA @Berkeley_AI @CMU_Robotics @StanfordAILab
capgym.github.io
🧵
English
Sandeep Bajamahal @ ICML🇰🇷 retweetledi
Human Archive builds datasets to model sensorimotor intelligence for robotics and world models.
HA-Multi is the largest dataset of manual labor tasks aligning egocentric RGB, stereo depth (active IR), tactile gloves, body IMUs, and wrist cameras.
Congrats on the launch, @babugi28, @shloke_patel, @samay_maini15, and @Rush26agarwal!
ycombinator.com/launches/PeP-h…
English
Excited to share our latest work and tools on scaling robot augmentation for generalist policies!
Check it out:
Lawrence Yunliang Chen@Lawrence_Y_Chen
Can robot data evolve with robot hardware? Robot arms and hands are rapidly evolving based on advances in materials, motors, sensors, mechanical designs, and applications. Now, demo data can also evolve: AugE-Toolkit is a new open-source, easy-to-use, scalable package for augmenting robot embodiments: it converts demo data from one robot arm/gripper to a different robot arm/gripper in a scalable way. We used AugE-Toolkit to create OXE-AugE––a large open-source dataset that augments 16 popular OXE subsets with 9 different robot embodiments to provide over 2M new trajectories, tripling the size of the original dataset. The OXE-AugE dataset is on HuggingFace and you can download and train right away! Most importantly, we find robot augmentation scales! 🌐 oxe-auge.github.io 🤗 huggingface.co/oxe-auge 🔗arxiv.org/abs/2512.13100 🧵👇 (1/9)
English
💻If you want to see the implementation and train a model yourself, the full PyTorch code is open-source.
Check out the repo here: github.com/sunnythecode/W…
#AI #MachineLearning #RL #Python #DeepLearning
English
Walking seems simple until you try to teach a robot to do it. 🤖
I used the Soft Actor-Critic (SAC) RL algorithm to train a bipedal walker. My deep dive covers the math and some findings from vision-based & stabilized approaches: @sunny.bajamahal/learning-to-walk-with-reinforcement-learning-191560eaa283" target="_blank" rel="nofollow noopener">medium.com/@sunny.bajamah…
Details threaded 🧵
English
💡The biggest win came from episode scaling. Gradually increasing episode duration prevented the critic network from overfitting to stationary data early on.
This allowed it to experience more crucial failures, stabilizing and speeding up training significantly. 📈
English
🧠 Standard Q-Learning struggles with the continuous action spaces needed for robotic joints.
SAC solves this by maximizing entropy (randomness) alongside reward. It encourages the robot to try weird movements (exploration) before settling on the best stride (exploitation).
English