What if you could train agents on a 𝗱𝗲𝗰𝗮𝗱𝗲 of driving experience in 𝘂𝗻𝗱𝗲𝗿 𝗮𝗻 𝗵𝗼𝘂𝗿, on a single GPU? Excited to share 𝙋𝙪𝙛𝙛𝙚𝙧𝘿𝙧𝙞𝙫𝙚 2.0: A fast, friendly driving simulator with RL training via PufferLib at 𝟯𝟬𝟬𝗞 𝘀𝘁𝗲𝗽𝘀/𝘀𝗲𝗰 🐡 + 🚗 youtu.be/LfQ324R-cbE?...

Our new E2E driving method, TransFuser v6, is out on ArXiv. It outperforms all other methods on CARLA by a wide margin, 95 DS on Bench2Drive! We show that minimizing the asymmetry between data annotator and policy is key for strong IL results. Code, models, and paper: ln2697.github.io/lead/

1/n REGLUE Your Latents! 🚀 We introduce REGLUE: a unified framework that entangles VAE latents ➕ Global ➕ Local semantics for faster, higher-fidelity image generation. Links (paper + code) at the end👇

Winter sea

Weekend views

Ok, I have a strong opinion on this, perhaps you can convince me otherwise: if you need an LLM to understand a paper for reviewing, perhaps you should not be a reviewer. And your argument is saved time, then perhaps you planned to not read all the details. Then again, you should not be a reviewer.

Full house for Gianni Franchi’s viva / HDR defense “Towards Trustworthy AI”. Forza Gianni!

Happy to announce that our paper "IPA: An Information-Reconstructive Input Projection Framework for Efficient Foundation Model Adaptation" got the best paper award at the #CCFM workshop at #NeurIPS2025 Many thanks to the reviewers and organizers. Kudos to @yuanyinnn.bsky.social & team!

Our @spyrosgidaris.bsky.social is speaking this morning (Wed, Dec 10th, 11:00 am Paris time) about "Latent Representations for Better Generative Image Modeling" in the Hi! PARIS - ELLIS monthly seminar. The talk will be live-streamed: www.hi-paris.fr/2025/09/26/a...

Robotics researchers, this is an occasion not to be missed: top-notch speakers in an excellent location in the Alps this February 🤖⛷️ It's the Winter School on Social Robotics, Artificial Intelligence and Multimedia (SoRAIM), 9-13 Feb 2026 👇

Tired Europe: Let's do tons of AI regulations Wired Europe: Let's do tons of AI open source #aiPULSE2025

I'm speaking at #aiPULSE2025 today on Open & re-purposable foundation models for the automotive industry. The morning keynotes talked a lot about open source so my slide here might be timely.

Hey #NeurIPS2025! We have a fun deck of works with the team and friends to present this year. Check them out!

We fermented our thoughts on understanding LoRA & ended up with IPA🍺 We found an asymmetry in LoRA: during training, A changes little & B eats most task-specific adaptation. So we pre-train A to preserve information before adaptation w/ excellent parameter efficiency #NeurIPS2025 #CCFM 👇

I also think that stopping the process and letting ACs decide based on original reviews would be the safest thing to do at this point to avoid any skewed or suspicion of skewed decisions because of the leak #ICLR2026

This is by a large margin the most serious problem and mistake in conference peer review I have seen in my career. Apparently a many people were aware of this and many could find out who their reviewers were. This probably has created a large number of unnecessary enmities. @iclr-conf.bsky.social

Check out NAF: an effective ViT feature upsampler to produce excellent (and eye-candy) pixel-level feature maps. NAF outperform both VFM-specific upsamplers (FeatUp, JAFAR) and VFM-agnostic methods (JBU, AnyUp) over multiple downstream tasks 👇

We have a few fun papers at #BMVC2025 Check them out!

Honored to host a very special guest today: Huy V. Vo, PhD alumnus of @valeoai.bsky.social He presented his work on automatic data-curation strategies for self-supervised representation learning (DINOv2, DINOv3). Find out more about his research here: huyvvo.github.io

Thrilled to have Jayneel Parekh visiting us over a break in the CVPR sprint to talk about his exciting works in analyzing MLLMs representations. Check out his great work jayneelparekh.github.io

Excellent speaker lineup for the @naverlabseurope.bsky.social AI for Robotics Workshop. For those at home, the event is live-streamed on the landing page: europe.naverlabs.com/updates/ai4r...

Pro-tip: if you send me an email asking for an internship, a PhD or a postdoc position, don't copy/paste an over-hyped summary of one of my papers you just asked ChatGPT to spit out. Pro-tip#2: don't do that with other professors as well. It's not just me.

#KostasThoughts: Please stop using AI tools to write reviews. I want your authentic thoughts, even if they contain spelling, grammar, & factual mistakes. An LLM should not influence your judgment. The purpose of multiple reviewers is to provide diverse views and error checking.

All the slop review drama unfolding at ICLR is painful to watch. But having so many eyes on it puts us in a better point to find solutions. This would have taken longer to be exposed on a closed platform where we may have had a few screenshots and a few author stories at best.

Post-CVPR deadline run, at last

Coarse plan for a 1-week rebuttal period: day 1: read the reviews day 2-6: calm down day 7: write the rebuttal

Unlike ICLR reviews, the sunrise never disappoints

My first smartphone was an iPhone 3GS and then I changed to a series of Android phones in 2013 (HTC One, Lenovo K6, Xiami Note 7, Google Pixel 6) but was always nostalgic about iOS. Now I switched back to an iPhone 16. I waited 2 weeks before writing this, but I am __VERY__ unhappy about iOS. 🧵 1/8

I recently had to switch to an iphone after an entire life on android phones. First shock is how bad and limited is the ios clock app compared to the google one -- I might be using alarms in very peculiar ways 😅 If you're out of ideas you can always have a look on what other folks are trying out.

We introduce MIRO: a new paradigm for T2I model alignment integrating reward conditioning into pretraining, eliminating the need for separate fine-tuning/RL stages. This single-stage approach offers unprecedented efficiency and control. - 19x faster convergence ⚡ - 370x less FLOPS than FLUX-dev 📉

The perks of early risers

Launching the Physical AI AV Dataset! 🚀 huggingface.co/datasets/nvi... One of the largest, most diverse & commercially usable open-source datasets for AVs. - 1727 hours of driving data - Camera, LiDAR, & radar - 25 countries, 2500+ cities This is just the beginning, more features to come!

This is a major reset for autonomous driving research! So exciting 🚀

X-mas came earlier this year! Nvidia has just released the huge Physical AI AV Dataset - 1727 hrs of driving data: 310K clips of 20s - sensor rig: 7 cameras, lidar, radar - 25 countries, 2.5K cities from US + Europe Kudos to @kashyap7x.bsky.social et al.! huggingface.co/datasets/nvi...

EuroHPC wrote a piece about our love of GPUs 😊 A EuroHPC Success Story | Clear Vision for Self-Driving Cars www.eurohpc-ju.europa.eu/eurohpc-succ...

A reminder which might be relevant now: we are looking to hire a senior research scientist in Robotics at @naverlabseurope.bsky.social in Grenoble, France.

Happy to represent Ukraine at #ICCV2025 . Come see my poster today at 11:45 (#399)!

Today is the day for our @iccv.bsky.social workshop that we've been cooking with excitement in the past months. Learning to See: Advancing Spatial Understanding for Embodied Intelligence #ICCV2025 Join us! opendrivelab.com/iccv2025/wor...

Trying to mitigate ICCV FOMO today. Staying at home this time. Have fun everybody!

Our recent research will be presented at @iccv.bsky.social! #ICCV2025 We’ll present 5 papers about: 💡 self-supervised & representation learning 🌍 3D occupancy & multi-sensor perception 🧩 open-vocabulary segmentation 🧠 multimodal LLMs & explainability valeoai.github.io/posts/iccv-2...

Belle lune

Hear! Hear!

“Has anyone heard about DUSt3R?” All hands and hearts up in the room. Honored to welcome @gabrielacsurka.bsky.social today to speak about the amazing work @naverlabseurope.bsky.social towards 3D Foundation Models

The unreasonable effectiveness of mastering the skill of placing <EOS> in real life

The outstanding reviewer award is definitely one of my favorite type of awards as good reviewing helps so much to advance our community and research forward. Congrats to all great reviewers out there whether rewarded this time or not! #iccv2025

Congratulations to our lab colleagues who have been named Outstanding Reviewers at #ICCV2025 👏 Andrei Bursuc @abursuc.bsky.social Anh-Quan Cao @anhquancao.bsky.social Renaud Marlet Eloi Zablocki @eloizablocki.bsky.social @iccv.bsky.social iccv.thecvf.com/Conferences/...

Some night runs seem more like sightseeing than running

Here we go again 😅 This time I’m planning to take a more senior role to help mentor the next gen of publicity chairs. Please consider volunteering!

Many treat uncertainty = a number. At Apple, we're rethinking this: LLMs should output strings that reveal all information of their internal distributions. We find that Reasoning, SFT, CoT can't do it - yet. To get there, we introduce the SelfReflect benchmark. arxiv.org/pdf/2505.20295