antoine dedieu
@antoine_dedieu
Followers
230
Following
369
Media
22
Statuses
141
Senior Research Scientist @GoogleDeepMind. Ex @vicariousai (acquired by Alphabet) @MIT ORC and @Polytechnique alumni 🇫🇷. Husband & Dad.
Joined May 2020
Happy to share our new preprint “Improving Transformer World Models for Data-Efficient RL”: We propose a ladder of improvements to model-based RL and achieve for the first time a superhuman reward on the challenging Craftax-classic benchmark!. 1/10
4
17
115
RT @AlecStapp: Still think about this every time I hear someone romanticizing the pre-industrial era
0
325
0
RT @jparkerholder: Genie 3 feels like a watershed moment for world models 🌐: we can now generate multi-minute, real-time interactive simula….
0
555
0
These days, my @Waymo commute isn't just a ride to work, it's also a photo op for tourists. Glad I could contribute to your vacation memories!.
1
0
9
Our paper is now published at ICML 🇨🇦: We improved our SOTA results on Craftax-classic (69.66%), Craftax-full (7.20%) and showed that our model-based RL ladder of improvements generalizes to Minatar environments. I'll be presenting our work later today.
Happy to share our new preprint “Improving Transformer World Models for Data-Efficient RL”: We propose a ladder of improvements to model-based RL and achieve for the first time a superhuman reward on the challenging Craftax-classic benchmark!. 1/10
1
1
7
Exciting news! Nuclear fusion has the potential to be a game changer in the climate crisis 🌎.
Historic news for CFS: We’ve signed a landmark agreement with @Google, a multifaceted partnership that dramatically advances our commercial fusion energy mission:. ⚡️Google signed an offtake agreement for 200 megawatts of power from our first ARC power plant.⚡️Google is
0
1
9
A remarkable achievement! .Thrilled to see @DarrenAronofsky, a director I have long admired, embracing AI to enhance human creativity and the movie-making process 🎥.
🎬 @DarrenAronofsky x @Eliza_McNitt present ANCESTRA: a true milestone in the history of AI cinema. Our latest collaboration at @GoogleDeepMind DeepMind with @primordialsoup_ & @Google Creative Lab. Ancestra combines live action and generative AI, leading the way in how we
0
0
6
RT @EmmanuelMacron: Paris, Europe’s tech capital and the world’s fourth-largest hub for its ecosystem. Make it iconic, choose France! http….
0
473
0
Unbelievable 🤯.
My first Veo 3 gen. > a video with dialogue of two muffins while baking in an over, the first muffin says "I can't believe this Veo 3 thing can do dialogue now!", the second muffin says "AAAAH, a talking muffin!"
0
0
3
A clean sweep during NBA playoffs is great timing 🧹🚀.
🚨Breaking: @GoogleDeepMind’s latest Gemini-2.5-Pro is now ranked #1 across all LMArena leaderboards 🏆. Highlights:.- #1 in all text arenas (Coding, Style Control, Creative Writing, etc).- #1 on the Vision leaderboard with a ~70 pts lead!.- #1 on WebDev Arena, surpassing Claude
0
0
1
AI is unlocking massive value by revolutionising archaic industries 🚀.
Secret's out! . Codi just dropped the 1st AI Office Assistant. Furniture? Cleaning? Beverages? Building issues? . Our AI office assistant handles it all for you. Just ask and it will get it done. ANY space. 4x more cost-effective (vs EA or FTE or founder's time. )
0
0
1
My colleagues @joeaortiz, @swaroopgj and @sirbayes will be in Singapore 🇸🇬 to present our recent work at the exciting World Models Workshop!.Do not hesitate to swing by and ask them questions.
🚨 Demos @ World Models Workshop #ICLR2025 🚨.👉 A glimpse into some of the demos from accepted papers at the workshop:. 🧠 Transformer World Models — boosting data-efficient RL @antoine_dedieu .🔄 VLM-in-the-loop Steering — align policies via latent
0
2
12
RT @demishassabis: Honoured to be on the #TIME100 list and cover - and congratulations to all the other people featured. Thank you to the b….
0
151
0
Our PGMax paper is out in @JmlrOrg! 🎉. Explore our JAX package for effortlessly building discrete PGMs and running scalable Loopy Belief Propagation:. 🔗 Paper: 💻 Code:
0
0
4
RT @danijarh: Excited to share that DreamerV3 has been published in Nature!. Dreamer solves control tasks by imagining the future outcomes….
0
152
0
RT @OfficialLoganK: Gemini 2.5 Pro, available to all users in @cursor_ai! . It’s time to build 🚢.
0
230
0
RT @MatthewBerman: Gemini 2.5 Pro is insane at coding. It's far better than anything else I've tested. Below are ONE SHOT demos 🧵 https:/….
0
313
0
RT @sirbayes: I'm happy to announce that v2 of my RL tutorial is now online. I added a new chapter on multi-agent RL, and improved the sect….
arxiv.org
This manuscript gives a big-picture, up-to-date overview of the field of (deep) reinforcement learning and sequential decision making, covering value-based method, policy-gradient methods,...
0
297
0