←
Efstratios Gavves954 posts
Efstratios Gavves avatar

Efstratios Gavves

@egavves

Associate Professor & Co-Founder - Dynamical Deep Learning

◷ Joined September 2011▧ 81 Photos and videos
Cees Snoek avatar
⟳ Efstratios Gavves retweeted
Cees Snoek@cgmsnoek·Oct 6

🛗 Introducing LiFT: A flow matching transformer that loops at inference far beyond its training depth, with no retraining, early exits, or other modifications. 💡

♾️ LiFT: Loop Flow Transformers 🧵👇 Diffusion models usually spend more inference compute by taking more sampling steps. What if we could also spend compute inside each step? LiFT turns part of a DiT into a recurrent core, letting us scale depth at inference with only a small architectural change. On ImageNet, it beats dense DiT-XL/2 with 3.34 lower FID, while using 60% fewer parameters, 32% fewer training FLOPs, and 52% fewer inference FLOPs. Jointly led with Mohammad Mahdi Derakhshani 👑(@mmderakhshani) And amazing supervisors Cees Snoek (@cgmsnoek), Jan-Willem van de Meent (@jwvdm) and Gertjan J. Burghouts (@gjburghouts)

1:29
Efstratios Gavves avatar
Efstratios Gavves@egavves·Jun 5

Great article, this line of work of “learnable digital twins”/“learnable simulators” is what we first pioneers last year with DreMa 😃 And there is more coming 😃 dreamtomanipulate.github.io/

Fei-Fei Li @drfeifei Jun 3 Article A Functional Taxonomy of World Models “The world is everything that is the case.” — Ludwig Wittgenstein, Tractatus Logico-Philosophicus, 1921 The world is not made of words. In an earlier essay, we argued that spatial intelligence is AI’s

Efstratios Gavves avatar
Efstratios Gavves@egavves·May 6

From World Models To Robot World Models!! Please come join us into our next workshop held in RSS in Sydney! #physicalai #worldmodels

🚀 Announcing #RSS2026 workshop: Robot World Models 🌐🤖🐨🌐 July 17, 2026 in 🇦🇺! We're bringing together researchers building the next generation of WM for robots — models robots can act in: controllable, long-horizon & grounded in real dynamics. 🔗 robot-worldmodels.github.io

Post attachment
Andrii Zadaianchuk 🇺🇦 avatar
⟳ Efstratios Gavves retweeted
Andrii Zadaianchuk 🇺🇦@ZadaianchukML·May 5

🚀 Announcing #RSS2026 workshop: Robot World Models 🌐🤖🐨🌐 July 17, 2026 in 🇦🇺! We're bringing together researchers building the next generation of WM for robots — models robots can act in: controllable, long-horizon & grounded in real dynamics. 🔗 robot-worldmodels.github.io

Post attachment
Ethan Mollick avatar
⟳ Efstratios Gavves retweeted
Ethan Mollick@emollick·2 Sep 2025

60 years of exponential growth in chip density was achieved not through one breakthrough or technology, but a series of problems solved and new paradigms explored as old ones hit limits. I don't think current AI has hit a wall, but even if it does, there many paths forward now.

Post attachment
Efstratios Gavves avatar
Efstratios Gavves@egavves·3 Aug 2025

The next wave of AI will be cyberphysical, embodied, empowered with robot worlds! Super psyched for our first workshop on robot world models, marrying Computer Vision, Robot Learning, and Deep Learning, and with a dream team of co-organizers including Toyota Robotics.

🚀 We’re excited to announce our #CoRL2025 workshop: Learning to Simulate Robot Worlds Spanning high-fidelity simulators, digital twins, and learned world models - our goal is to unite communities to push robot learning forward 🤖🌐 🔗 simulatingrobotworlds.github… 🧵🧵🧵

Andrii Zadaianchuk 🇺🇦 avatar
⟳ Efstratios Gavves retweeted
Andrii Zadaianchuk 🇺🇦@ZadaianchukML·1 Aug 2025

🚀 We’re excited to announce our #CoRL2025 workshop: Learning to Simulate Robot Worlds Spanning high-fidelity simulators, digital twins, and learned world models - our goal is to unite communities to push robot learning forward 🤖🌐 🔗 simulatingrobotworlds.github… 🧵🧵🧵

The Greek Analyst avatar
⟳ Efstratios Gavves retweeted
The Greek Analyst@GreekAnalyst·18 Jul 2025

I have not vouched for or promoted any other event in 10 years of existing on twitter. I broke my policy for this one. It’s not the AI part that makes this conference unique. It’s the collective brain power that is off the charts.

If you are in Athens this weekend, there is only one place to be (and no, I’m not referring to the beach). Greeks in AI 2025, the annual gathering of the most amazing Greek AI scientists and practitioners, will take place on 19-20 July. An event not to be missed!

Post attachment
Efstratios Gavves avatar
Efstratios Gavves@egavves·29 Jun 2025

Or later ICCV2025 on temporal self Sorenson, check it out!

New paper out - accepted at @ICCVConference We introduce MoSiC, a self-supervised learning framework that learns temporally consistent representations from video using motion cues. Key idea: leverage long-range point tracks to enforce dense feature coherence across time.🧵

Post attachment
Shashank avatar
⟳ Efstratios Gavves retweeted
Shashank@shawshank_v·26 Jun 2025

New paper out - accepted at @ICCVConference We introduce MoSiC, a self-supervised learning framework that learns temporally consistent representations from video using motion cues. Key idea: leverage long-range point tracks to enforce dense feature coherence across time.🧵

Post attachment
Yuki avatar
⟳ Efstratios Gavves retweeted
Yuki@y_m_asano·26 Jun 2025

New paper accepted at @ICCVConference: MoSiC, our new post-pretraining technique for upgrading vision foundation models like DINOv2R using videos, thanks to strong point trackers and Sinkhorn clustering. Check the thread below :)

New paper out - accepted at @ICCVConference We introduce MoSiC, a self-supervised learning framework that learns temporally consistent representations from video using motion cues. Key idea: leverage long-range point tracks to enforce dense feature coherence across time.🧵

Post attachment
Efstratios Gavves avatar
Efstratios Gavves@egavves·13 Jun 2025

An excellent work by @ZadaianchukML , it's not every year that an MSc project gets a CVPR oral!!

How to represent dynamic real-world data both consistently and efficiently, while reflecting the compositional object-centric structure of the world? Contrast your slots! ...with our new SlotContrast method(🚀#CVPR2025 Oral🚀)! 🌐website: slotcontrast.github.io 🧵🧵🧵 1/n

0:04