Christian Wolf (🦋🦋🦋) (@chriswolfvision) 's Twitter Profile
Christian Wolf (🦋🦋🦋)

@chriswolfvision

Principal Scientist, @NaverLabsEurope, Lead of Spatial AI team. AI for Robotics. IEEE-PAMI area editor. Feedback: admonymous.co/chriswolfvision

ID: 4249537197

linkhttps://chriswolfvision.github.io/www calendar_today22-11-2015 12:41:47

5,5K Tweet

7,7K Followers

746 Following

Austrian Airlines (@_austrian) 's Twitter Profile Photo

Dear guests, you can stay connected with us on Instagram, TikTok, Facebook & LinkedIn. Feel free to reach out anytime via our Austrian App or at austrian.com.

Dear guests, you can stay connected with us on Instagram, TikTok, Facebook & LinkedIn. Feel free to reach out anytime via our Austrian App or at austrian.com.
Christian Wolf (🦋🦋🦋) (@chriswolfvision) 's Twitter Profile Photo

So what is the nettiquette with cross-posting? Can I take a screen shot of a conversation with contacts on one platform and post it on the other one, or would that be a breach of trust? Assuming no malicious intent or sarcasm or something of that sort, of course.

Christian Wolf (🦋🦋🦋) (@chriswolfvision) 's Twitter Profile Photo

Editing a video with iMovie is like riding a tricycle: do it at the first try, no need for explanations. Switching to Final Cut Pro is like landing an F-14 on an aircraft carrier, at night, without a pilot's license. With the ILS broken and the afterburner stuck on "On".

Christian Wolf (🦋🦋🦋) (@chriswolfvision) 's Twitter Profile Photo

There are now several benchmarks testing spatial reasoning and agent capabilities of LLMs and VLMs: - arxiv.org/abs/2410.06468 (does spatial cognition ...) - arxiv.org/abs/2307.06281 (MMBench) - arxiv.org/abs/2411.13543 (BALROG) - additional points for the LOTR ref.

There are now several benchmarks testing spatial reasoning and agent capabilities of LLMs and VLMs:  

- arxiv.org/abs/2410.06468 (does spatial cognition ...) 
- arxiv.org/abs/2307.06281 (MMBench)
 - arxiv.org/abs/2411.13543 (BALROG) - additional points for the LOTR ref.
Christian Wolf (🦋🦋🦋) (@chriswolfvision) 's Twitter Profile Photo

Our e2d trained navigation agent has been accepted to #CVPR2025: fast-moving, efficient: correct modelling of motion during training in simulation. arxiv.org/abs/2503.08306 By S Janny, H Poirier, L Antsfeld, G Bono, G Monaci, B Chidlovskii, F Giuliari, A Del Bue and myself.

Christian Wolf (🦋🦋🦋) (@chriswolfvision) 's Twitter Profile Photo

Interactive website: europe.naverlabs.com/research/publi… Large-scale study w 262 real episodes finds evidence for presence of realistic dynamics used for open-loop forecasting, interplay with sensing; usage of latent memory (emerging maps); value relates to long-term plans. #CVPR2025

Christian Wolf (🦋🦋🦋) (@chriswolfvision) 's Twitter Profile Photo

We find evidence for a plan structured on the level of paths; that its estimate of success goes beyond the effect of the next action. Abandoning a navigation option for a more promising one increases the value estimate, as the agent now expects a higher future return. #CVPR2025

Christian Wolf (🦋🦋🦋) (@chriswolfvision) 's Twitter Profile Photo

A big thanks to the NAVER LABS Europe Spatial AI team for this, but in particular to Steeven Janny, not only for the great work linking e2e training to a Kalman filter but also the D3.js magic making the real-time analytics work, and the interactive website!

Christian Wolf (🦋🦋🦋) (@chriswolfvision) 's Twitter Profile Photo

We have written a blog post on our research in end-to-end trained models for navigation: - Fast and precise motion by training w realistic motion models - Geometric foundation models - Efficient training losses. europe.naverlabs.com/research/found… (The NAVER LABS Europe Spatial AI team)

We have written a blog post on our research in end-to-end trained models for navigation:

- Fast and precise motion by training w realistic motion models
- Geometric foundation models
- Efficient training losses.

europe.naverlabs.com/research/found…

(The <a href="/naverlabseurope/">NAVER LABS Europe</a>  Spatial AI team)
Christian Wolf (🦋🦋🦋) (@chriswolfvision) 's Twitter Profile Photo

We have made some post-conference improvements of our CVPR'25 paper on end-to-end trained navigation. The agent has similar success rate but is more efficient, faster, less hesitant. Will be presented at CVPR in June. arxiv.org/abs/2503.08306 Work by Steeven Janny et al.