Reinforcement Learning Researcher at Agile Robots and @UniFreiburg. Previously @BoschGlobal, @RWTH, @TMDTWuppertal.
Ulm, Germany
Joined November 2017
- Tweets283
- Following827
- Followers259
- Likes2.8K
Moritz Schneider retweeted
Behavioral cloning mystery
seohong.me/blog/behavioral-c…
I wrote a new blog post about "mysteries" in behavioral cloning that appear with real-world robot data (e.g., overfitting is "good"). I also tried to demystify them and shared my thoughts!
Moritz Schneider retweeted
Action chunking is a critical component in virtually all modern approaches to imitation learning for robotics.
But why is it so critical, and do we really need action chunking? Check out our latest work to find out! (1/n)
action-chunking.github.io
Moritz Schneider retweeted
Revenue stonks among robotics companies - finally! 📈
Agile Robots doubling revenue to €600M this year!
German robotics company Agile Robots hit €300 million ($346M) in revenue last year and expects to double that to €600 million in 2026.
Their CEO, Zhaopeng Chen told the Wall Street Journal they've already signed customer contracts backing those numbers. Profitability within 2-3 years.
Founded in Munich in 2018, Agile Robots raised $1.5 billion from SoftBank and others. They're not just an AI company, because they own the manufacturing stack, and that's the real difference.
They've bought over a dozen companies. thyssenkrupp Automation Engineering. Idealworks. Franka Robotics, BÄR Automation GmbH, many of them ;)
10-15% of new business this year comes through acquisitions. They're acquiring capability, customer bases, manufacturing expertise.
Recently partnered with Google DeepMind to integrate Gemini Robotics into their systems. AI foundation models plus industrial automation hardware in one package.
Oh boy, it hits hard for me to start writing about doubling the revenue numbers in the robotics space.
WHAT A TIME TO BE A ROBOTICS GUY! 😮💨
Read more (paywalled tho): wsj.com/tech/ai/german-robot…
~~
♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com
I am really excited to finally see what DeepMind creates using our robots.
Looking forward to see more results and experiments!
Moritz Schneider retweeted
3D Scene Graphs are becoming popular for robot navigation, manipulation & LLM-based scene reasoning.
But what is a 3DSG, how are they built, and what's still missing? Read our survey for the answers!
📕 arxiv.org/abs/2606.19383
🌐 3dscenegraphs.com Searchable 3DSG database
Moritz Schneider retweeted
World-Action Models (WAMs) have become the second dominant recipe for robot foundation models, next to classical VLAs.
So where do they come from, and how do they compare vs VLAs?
I wrote an small overview of the WAM landscape, with some personal takes:
developer.nvidia.com/blog/pr…
Moritz Schneider retweeted
summoning the design master @GoingBallistic5 😮💨
Moritz Schneider retweeted
Have you seen Agile Robots’ latest humanoid robot? 🤖
Hear from Laurenz von Czettritz as he introduces Agile ONE, a coworker on the factory floor, and explains how NVIDIA technology powers its advanced capabilities.
Moritz Schneider retweeted
Excited to partner with Agile Robots! Looking forward to seeing our models being deployed through Agile Robots incredible platform to help solve some of the most complex industrial challenges
Google DeepMind 🤝 Agile Robots
Our new research partnership will integrate the Gemini foundation models with their hardware to help build the next generation of more helpful and useful robots.
Find out more → goo.gle/4lKu7de
Moritz Schneider retweeted
Google DeepMind 🤝 Agile Robots
Our new research partnership will integrate the Gemini foundation models with their hardware to help build the next generation of more helpful and useful robots.
Find out more → goo.gle/4lKu7de
Moritz Schneider retweeted
NVIDIA GTC 2026 wrapped up after an inspiring week at the intersection of AI & robotics. With Agile Robots, we showcased FR3 Duo running an assembly task with @nvidia's GR00T N1.6 model—no scripting, real-time perception to motion.
A standout moment: Jensen Huang signing the Agile ONE humanoid.
#NVIDIAGTC2026
Moritz Schneider retweeted
World Modeling research needs fast iteration, reproducibility, optimized baselines, open-source, and precise zero-shot stress testing. Here comes stable-worldmodel!
Paper: arxiv.org/abs/2602.08968
Code: github.com/galilai-group/sta…
Come stress-test your model/idea!
DINO-WM results ⬇️
Moritz Schneider retweeted
📣 UNITE: Unified Semantic Transformer for 3D Scene Understanding 📣
Given only a couple of images of a scene, UNITE jointly recovers:
✅ 3D Scene Geometry
✅ Semantic Segmentation ✅ Instance Segmentation
✅ Object Articulations
🚀 Project page: unite-page.github.io/
Moritz Schneider retweeted
Excited to introduce Dreamer 4, an agent that learns to solve complex control tasks entirely inside of its scalable world model! 🌎🤖
Dreamer 4 pushes the frontier of world model accuracy, speed, and learning complex tasks from offline datasets.
co-led with @wilson1yan
Moritz Schneider retweeted
The 🔥Hype🔥 is real 🤩 #OpenHype got accepted at #NeurIPS2025
TL;DR: We encode CLIP features in a hyperbolic space, where object hierarchies are naturally encoded along geodesic paths, making it easy to reason about parts of objects in 3D.
👨💻Website: lisaweijler.github.io/openhy…
Moritz Schneider retweeted
Excited to share that DreamerV3 has been published in Nature!
Dreamer solves control tasks by imagining the future outcomes of its actions inside of a continuously learned world model 🌏
It's the first agent to find diamonds in Minecraft from scratch without human data! 💎
👇
Moritz Schneider retweeted
🎉 Excited to share RelationField: Relate Anything in Radiance Fields, accepted at #CVPR2025
The first method to extract inter-object relationships directly in radiance fields.
📝Paper arxiv.org/abs/2412.13652
🚀Project relationfield.github.io
👨💻Code github.com/boschresearch/Rel…
I will be at #NeurIPS2024 this week, presenting our paper "The Surprising Ineffectiveness of Pre-Trained Visual Representations for Model-Based Reinforcement Learning" on Wednesday morning.
📝OpenReview openreview.net/forum?id=LvAy…
🖥️schneimo.com/pvr4mbrl/
TL;DR: Our results challenge some common assumptions about the benefits of PVRs in MBRL and highlight the importance of data diversity and reward prediction accuracy.