Large-Scale Robot Decision Making @GoogleDeepMind European @ELLISforEurope - imitation interaction transfer - priors: @oxfordrobots @berkeley_ai @ETH @MIThttps://bsky.app/profile/mwulfmeier.bsky.socialhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3mh6lsb5tkk2nReview request: As usual for the time of year, I'll be looking for #IROS2026 reviewers. Highly interesting stack of papers on #ReinforcementLearning #Sim2Real #RewardLearning #LLMs #DataEfficiency #RoboticManipulation Reach out with your ID or papercept registered mail address and background.16 Mar 2026 13:46 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3mh6lsb5tkk2nhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3mdgbfbpptk2uRobotiq directly integrating touch, robustly and reliably (hopefully 😉), is a big deal. looking forward to playing with these. Now reannotating old teleop data might be unfeasible, but RL... 🚀 https://www.linkedin.com/posts/vision-gets-robots-to-contact-touch-ugcPost-7421909036687810560-a3_g?utm_source=share&utm_medium=member_android&rcm=ACoAAA7d1W8BpzOeQTJ9Z9DS-ektiobmtblIVUg27 Jan 2026 17:20 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3mdgbfbpptk2uhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3mcv2qjkpe22nNice perspective from Lu Li, Tianwei Ni, Yihao Sun, and Pierre-Luc Bacon on different conditions for offline-to-online learning. lnkd.in/dUKc4jyV We have been heavily relying on replay across experiments (lnkd.in/duVpCUru). Few experiments get started these days without initial datasets available. https://lnkd.in/dUKc4jyV20 Jan 2026 21:06 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3mcv2qjkpe22nhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3mcv2pnzj4c2nDistributional RL has been the foundation of our recent projects (e.g. lnkd.in/eYiCp-y8). Nice to see a distributional perspective to improvement for LLMs. Amin Rakhsha,Kanika Madan,Tianyu Zhang, Amir Farahmand, and Amir Khasahmadi via mode estimation rather than best of N for LLM: lnkd.in/d4NpyegJ https://lnkd.in/eYiCp-y820 Jan 2026 21:05 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3mcv2pnzj4c2nhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3mccvvo2hjs2dCurriculum matters more for active data generation than for fixed datasets. Data defines what you can learn and RL curriculum enables finding different datasets. Great to see our paper mentioned in @dwarkesh_sp’s blog post on RL efficiency. dwarkesh.com/p/bits-per-sam… arxiv.org/pdf/1707.05300 https://dwarkesh.com/p/bits-per-sam…13 Jan 2026 15:51 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3mccvvo2hjs2dhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3mbwq27cl5c2hDidn't spend much time here and just had a look back: since when is most posting on bsky so aggressive and vile? And why? How can this get worse than x?08 Jan 2026 19:34 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3mbwq27cl5c2hhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3mbu463nofs2hWhile the human form is brilliant, it’s ultimately limited. The real future of robotics lies in the permission to move beyond our own biological constraints and optimize. Note, we didn't achieve supersonic flight by mimicking the flapping of wings.07 Jan 2026 18:33 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3mbu463nofs2hhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3mbtpi4jocc23Taking parental leave for some deeply needed reading and it's fascinating to see the walls between SFT and RL crumbling. 🧵07 Jan 2026 14:46 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3mbtpi4jocc23https://bsky.app/profile/mwulfmeier.bsky.social/post/3m73u3p4q6c2tJoin our NeurIPS social event: The Role of AI in Scientific Peer Review. lnkd.in/dgdBiBqM Help build community and explore solutions for a fair, efficient, and transparent peer review system.           Wed. Dec. 3rd, 7:00 PM – 9:00 PM (Upper Level Ballroom 6CDEF) https://lnkd.in/dgdBiBqM03 Dec 2025 16:14 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3m73u3p4q6c2thttps://bsky.app/profile/mwulfmeier.bsky.social/post/3m6wibl2ajk2iProgress on robotics and AI has never been more closely linked. Access to leading frontier models has the potential to define what's possible in the physical world. I'll be at #NeurIPS2025 for a couple of days next week. Find me if you want to know more. https://x.com/sundarpichai/status/1971244716046872577?t=Xrx-Rxaym4W-1OxOl5JebQ&s=0901 Dec 2025 13:00 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3m6wibl2ajk2ihttps://bsky.app/profile/mwulfmeier.bsky.social/post/3m6p35vcclk22I'll be at #NeurIPS2025 next week! Looking forward to catching up with old and new friends - on AI for robots, Gemini Robotics, the RLnaissance in AI, and bad puns. Bon appetit!28 Nov 2025 14:16 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3m6p35vcclk22https://bsky.app/profile/mwulfmeier.bsky.social/post/3m3mm22mlxc2dTruly enjoyed discussing the consolidation of specialist and generalist approaches to physical AI at #IROS2025. Hoping to visit Hangzhou in physical rather than digital form myself in the not too distant future - second IROS AC dinner missed in a row. #Robotics #physicalAI20 Oct 2025 10:25 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3m3mm22mlxc2dhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3m3mlyheikk2dThe explosion of AI capability and complexity demands better understanding. I firmly believe in studying large model behavior - from the perspective of artificial sequential decision making (like Inverse RL) to now linking it with human decision-making and cognitive models. bit.ly/3Wqrtxl http://bit.ly/3Wqrtxl20 Oct 2025 10:24 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3m3mlyheikk2dhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3m2kjfevohc2z'Papa, was ist das?' As a parent, Gemini has made my life massively easier. We found this caterpillar earlier and the answer is in fact correct (verification is much easier than generation thanks to classical Google search) Very curious about how LLMs are continuing to change the ways we learn!06 Oct 2025 21:07 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3m2kjfevohc2zhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lztweddozk2vThe position paper track at #NeurIPS2025 was a great idea, the acceptance rate of under 6% not so much! This is unnecessarily low and will reduce interest in any future iterations of the track. (Disclaimer: our team is part of the 94%) @NeurIPSConf27 Sep 2025 21:28 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lztweddozk2vhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lztubamluc2vIf you want one brain for any robot, cross-embodiment learning is the key. Check out the new model and tech report for sparks of it https://share.google/H5RBiwtCWnW7pvlWb And catch up with the team (unfortunately without me this year) at #CORL2025! More here: https://x.com/sundarpichai/status/1971244716046872577?t=BaBClx32X14OZeD2MzWDTQ&s=1927 Sep 2025 20:50 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lztubamluc2vhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lxzsp3ubrs2hControl problems are everywhere, but some applications of Reinforcement Learning are truly out of this world! 🌌 Our team's latest research @ Google DeepMind in #Science shows RL can improve sensitivity by 30-100x. See how RL can accelerate cosmic discovery!👇 (Image #nanobanana) lnkd.in/e8r7t3NJ04 Sep 2025 18:48 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lxzsp3ubrs2hhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3ltzsnmzmu22b#Robotics is hard, and so is #AutonomousDriving! Massive congratulations to my friends at #Waymo for proving me wrong! https://x.com/Waymo/status/1945106097741664630?t=c087Rd-GqZWhydv1kgDzJA&s=09 There was a time a couple years back when I was getting skeptical about the scale of both technological and societal challenges for scaling autonomy....15 Jul 2025 21:06 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3ltzsnmzmu22bhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lsm7q2fohk2bLarge language and vision models alone don't solve the whole #robotics problem. But they surely have a massive impact on generalisation and robustness! New scenes, new backgrounds, new objects, new people, new language and audio.... https://x.com/ayzwah/status/1937927586048577622?t=-9tt2mB2Iqf4qNjObqoh_A&s=1927 Jun 2025 17:58 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lsm7q2fohk2bhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lsecxounv22yDon't have a robot? Try our newest Gemini Robotics on-Device VLA in simulation! Or become a trusted tester and tune and adapt the model yourself! https://www.youtube.com/watch?v=nVMY3-kWhOc24 Jun 2025 14:34 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lsecxounv22yhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lsebil56ws2yThrilled to share something new: Fast, robust, and increasingly intelligent 🧠 🤖 Truly proud of our team and hoping you'll enjoy the model! https://x.com/GoogleDeepMind/status/1937511515768176966?t=7KebfEnimGzITlMzZmoeuA&s=1924 Jun 2025 14:08 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lsebil56ws2yhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lrvl3bmpoc2rThe hardest problem with RL for robots is not RL!18 Jun 2025 17:50 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lrvl3bmpoc2rhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lrq6c2j4yc2eAre there any great summary posts from @huggingface 's #LeRobot hackathon? What works, what doesn't, how many robots broken ...? 🤖16 Jun 2025 14:17 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lrq6c2j4yc2ehttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lptxbrhvxc26Looking forward to talking about how much multiagentness do we need for multiagent RL for #robot soccer and other real world applications! In half an hour at the #Robolethics workshop #ICRA2025 @GoogleDeepMind23 May 2025 15:32 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lptxbrhvxc26https://bsky.app/profile/mwulfmeier.bsky.social/post/3lpmr3boqzk2jIf you blink, you've missed it! Demis Hassabis showing a glance into the future of robotics and vision language action models! #GoogleDeepMind #GoogleIO #robotics20 May 2025 18:53 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lpmr3boqzk2jhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lpmhu46dx22qLooking forward to #ICRA2025 ! Find me (remotely) during the workshops and catch up on large-scale RL, IL, and multi agent learning with a side of vision language action models!20 May 2025 16:08 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lpmhu46dx22qhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lnuvjwfais23Agentic LLMs need to explore! The web is a non-stationary, highly partially unobservable 🏔️. Thomas #GoogleDeepMind internship systematically investigates current challenges and the role of scale, RL, and chain-of-thought reasoning to overcome them. arxiv.org/abs/2504.1607828 Apr 2025 13:43 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lnuvjwfais23https://bsky.app/profile/mwulfmeier.bsky.social/post/3lnl2poti5c2xVisiting #ICLR2025? Join for the 7th iteration of the #RobotLearning workshop 🤖🧠 (rather than the usual at NeurIPS)! Rather than human-like form, how far are we from human-level abilities? Join the fun in person (while I will be cheering the team from London)! www.robot-learning.ml/2025/ https://www.robot-learning.ml/2025/24 Apr 2025 15:49 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lnl2poti5c2xhttps://bsky.app/profile/mwulfmeier.bsky.social/post/3lmkg5dtmfs2zAs data becomes more valuable than compute, being able to invest compute to extract more information from the same amount of data is invaluable! https://youtu.be/6nJZopACRuQ?si=UnvUJpHpvEF5JOM_&t=552 Apprenticeship learning, inverse RL, ... are becoming really interesting again! https://proceedings.neurips.cc/paper_files/paper/2024/hash/a5036c166e44b731f214f41813364d01-Abstract-Conference.html11 Apr 2025 16:16 +0000at://did:plc:5bvf6odmcqp4ljmaprxilidu/app.bsky.feed.post/3lmkg5dtmfs2z