Michal Nauman (@mic_nau) header
Michal Nauman (@mic_nau) profile picture

Visiting scholar @ robot learning lab UC Berkeley. PhD student in ML/Robotics.

Joined February 2022
Photo by Michal Nauman (@mic_nau)Photo by Michal Nauman (@mic_nau)Photo by Michal Nauman (@mic_nau)Photo by Michal Nauman (@mic_nau)Photo by Michal Nauman (@mic_nau)Photo by Michal Nauman (@mic_nau)Photo by Michal Nauman (@mic_nau)Photo by Michal Nauman (@mic_nau)Photo by Michal Nauman (@mic_nau)Photo by Michal Nauman (@mic_nau)
Scaling has done wonders for deep learning, but for a long time it failed in on-policy RL... until now! We show that when done appropriately, scaling leads to state-of-the-art results in a variety of continuous control tasks🔥 Introducing BRO: Bigger, Regularized, Optimistic! 🧵
10
46
7
292
48,901
I’m hiring the first PhD students to establish the Zurich Embodied AI Lab at ETH, starting August 2027! We’ll work on general-purpose embodied intelligence for humanoid robots and dexterous manipulation. If you’d like to help shape the lab from day one, I’d love to hear from you. Application link below ⬇️
7
75
10
684
39,399
We found that value-based RL is a good fit for multi-task fine-tuning of VLAs. Check out our new paper linked below!
Can RL post-training of generalist robot policies enable improvement not just on the tasks we train on, but also generalization beyond them, improving on novel tasks not seen in training? We show that, with the proper RL recipe, we can enable such "generalist improvement"! 🧵
1
18
612
Michal Nauman retweeted
Game changer for humanoid full-body control training. Now it's superfast (and cheap, just one GPU) to train a single policy on *all* the motions. How come: (1) World model much faster than regular sim (2) Backprop through time collects way more signal than policy gradients
We found a way to scale humanoid control that reduces GPU-hours by more than 1000×. Across 129,514 motions, SuperMimic surpasses the state of the art (128 GPUs, 7 days) with 1 GPU in 3 hours. Project: supermimic.github.io Demo: supermimic.github.io/demo/
19
68
7
694
54,690
Honored to receive the 2026 Witold Lipski Award! Sincere thanks to my mentors, @marek_a_cygan and @pabbeel - this would not have happened without your support and guidance 🫡
Dwóch młodych badaczy #AI otrzymało Nagrodę im. Witolda Lipskiego, jedną z najbardziej prestiżowych w polskiej informatyce. Laureatami konkursu organizowanego przez Instytut Badawczy #IDEAS zostali: Hubert Baniecki @hbaniecki i Michał Nauman @mic_nau forumakademickie.pl/mlodzi-b…
1
2
49
1,737
Michal Nauman retweeted
A few years ago, learning robot learning meant stitching together dozens of papers and courses — with no clear path from the basics to what state-of-the-art systems actually do. This was one of the motivations behind creating @ETH's course "Robot Learning: From Fundamentals to Foundation Models", to provide a structured path from first principles all the way to modern foundation models for robotics. I strongly believe that education should be accessible to everyone, so I have made all lecture recordings publicly available on YouTube. Creating this course was one of the most challenging projects I have taken on. It was my first time designing and teaching an entire curriculum from scratch, while simultaneously working full-time in industry. On top of that, the course proved to be more popular than expected and we had to scale it to almost 300 students, which was only possible thanks to an amazing team of TAs. Looking back, it was an absolute privilege to teach this class and an incredibly rewarding experience. If you are getting into robot learning, this is the starting point I wish I had. 📚 Main lectures: youtube.com/watch?v=X0k14u6p… 🎤 Guest lectures: youtube.com/watch?v=aG8NPTPh… 🌐 Course website: cvg.ethz.ch/lectures/Robot-L…
20
66
7
500
47,362
Presented tomorrow at ICLR (Poster Session 5)!
🚀 Excited to share REPPO, a new on-policy RL agent! TL;DR: Replace PPO with REPPO for fewer hyperparameter headaches and more robust training. REPPO, led by @c_voelcker, will be presented at #ICLR2026. How does it work? 🧵👇
1
12
749
Michal Nauman retweeted
Growing as an RL researcher---like many others---my work mostly following the following template: 1) formulate the right problem setting 2) come up with an algorithm that solves it 3) show that it works in simulation like (the amazing!) MuJoCo.
Sim-to-online RL will be a key component to effectively achieving mastery in physical AI. In a massive empirical effort, Yarden As and the team did a fantastic job to systematically ablate design choices across 100+ real-world training runs on three distinct robotic platforms. arxiv.org/pdf/2602.20220v1 Relevant insights below 🧵👇 Yarden As, Dhruva Tirumala, Rene Zurbrugg, @breadli428, @StelianCoros, @arkrause @ETH @GoogleDeepMind
1
5
1
35
8,057
We have now released the code and wandbs for "What Does Flow-Matching Bring To TD Learning?" ! Code -- github.com/CMU-AIRe/floq/tre… Wandbs -- docs.google.com/spreadsheets… Feel free to reach out if you have any questions or issues! teapawt.net/aviral_kumar2/status/2…
🚨🚨 New paper on flow-matching value functions Last year, we showed training RL value functions with a flow-matching loss achieved SOTA results. But why does it work? And what could it possibly tell us about other things that have nothing to do with VFs or even RL? Short answer: iterative compute used correctly can address feature plasticity in continual learning! 🧵⬇️
1
17
149
15,107
Check out our new paper🔥 Here, we take a deeper look at why flow-matching works so well for modelling Q-values in RL. Surprisingly, we find that the iterative nature of flow matching improves plasticity of the learner, making it a great fit for value learning and RL in general
🚨🚨 New paper on flow-matching value functions Last year, we showed training RL value functions with a flow-matching loss achieved SOTA results. But why does it work? And what could it possibly tell us about other things that have nothing to do with VFs or even RL? Short answer: iterative compute used correctly can address feature plasticity in continual learning! 🧵⬇️
1
1
12
637
Michal Nauman retweeted
Can just a 4B model solve IMO-level proof problems at the level of much stronger LLMs like Gemini 3 Pro? Yes, if you can train the LLM to scale test-time compute well! We're very excited to release our 4B model "QED-Nano", built via an awesome open collab! Details below🧵⬇️
8
26
4
167
22,296
🚀 Excited to share REPPO, a new on-policy RL agent! TL;DR: Replace PPO with REPPO for fewer hyperparameter headaches and more robust training. REPPO, led by @c_voelcker, will be presented at #ICLR2026. How does it work? 🧵👇
3
21
1
120
14,270
llms sometimes lie even when they know the truth. activation steering can fix this, but where you intervene matters. we introduce gaussian depth scheduling: improves honesty in all 7 models (up to +17.2 points), training-free and model-agnostic. see: arxiv.org/abs/2512.07667
1
5
259
Interested in scaling Q-learning to billion parameter models and many tasks?🔥 Join us tomorrow in Exhibit Hall C,D,E #305 (4:30-7:30)! @pabbeel @aviral_kumar2 @marek_a_cygan @carlo_sferrazza
Multi-task RL can be highly sample-efficient and when done right, it unlocks LLM-style transfer and fine-tuning. We’re excited to introduce BRC, a simple recipe for multi-task RL that outperforms SOTA single-task agents while using less compute (!)
7
5
1
27
6,455
When we released HumanoidBench, I was excited about the prospect of a benchmark using a single embodiment for many different tasks. With BRC, we show that learning across such diverse tasks is significantly more effective than training on each task separately.
Multi-task RL can be highly sample-efficient and when done right, it unlocks LLM-style transfer and fine-tuning. We’re excited to introduce BRC, a simple recipe for multi-task RL that outperforms SOTA single-task agents while using less compute (!)
1
8
66
7,854
Check out our work, led by @mic_nau on scaling up multi-task RL using some simple, but very effective design choices! He will present this work at NeurIPS -- don't lose an opportunity to talk to him!!! 👇
Multi-task RL can be highly sample-efficient and when done right, it unlocks LLM-style transfer and fine-tuning. We’re excited to introduce BRC, a simple recipe for multi-task RL that outperforms SOTA single-task agents while using less compute (!)
2
32
5,525
Multi-task RL can be highly sample-efficient and when done right, it unlocks LLM-style transfer and fine-tuning. We’re excited to introduce BRC, a simple recipe for multi-task RL that outperforms SOTA single-task agents while using less compute (!)
4
27
4
114
29,536
I’ll be presenting this work next week at @NeurIPSConf - reach out to me if you’d like to chat or meet in San Diego! This project was done during my visit to BAIR Berkeley. Huge thanks to my amazing coauthors: @pabbeel @aviral_kumar2 @marek_a_cygan @carlo_sferrazza
1
1
12
731