Christian Wolf
@chriswolfvision.bsky.social
Principal Scientist at Naver Labs Europe, Lead of Spatial AI team. AI for Robotics, Computer Vision, Machine Learning. Austrian in France. https://chriswolfvision.github.io/www/
Dubai style chocolate is now also produced by Lindt.... and costs 15€ for 100g. What the....?
November 25, 2025 at 6:16 PM
Dubai style chocolate is now also produced by Lindt.... and costs 15€ for 100g. What the....?
Being happy about a lower trade deficit because of lower imports is like
- being happy of paying less income tax because of a lower salary, or
- not having to pay for gas anymore because the car is wrecked.
- being happy of paying less income tax because of a lower salary, or
- not having to pay for gas anymore because the car is wrecked.
November 24, 2025 at 12:56 PM
Being happy about a lower trade deficit because of lower imports is like
- being happy of paying less income tax because of a lower salary, or
- not having to pay for gas anymore because the car is wrecked.
- being happy of paying less income tax because of a lower salary, or
- not having to pay for gas anymore because the car is wrecked.
Reposted by Christian Wolf
🚗🌐 Working on domain adaptation for 3D point clouds / LiDAR?
We'll present MuDDoS at BMVC: a method that boosts multimodal distillation for 3D semantic segmentation under domain shift.
📍 BMVC
🕚 Monday, Poster Session 1: Multimodal Learning (11:00–12:30)
📌 Hadfield Hall #859
We'll present MuDDoS at BMVC: a method that boosts multimodal distillation for 3D semantic segmentation under domain shift.
📍 BMVC
🕚 Monday, Poster Session 1: Multimodal Learning (11:00–12:30)
📌 Hadfield Hall #859
November 24, 2025 at 5:00 AM
🚗🌐 Working on domain adaptation for 3D point clouds / LiDAR?
We'll present MuDDoS at BMVC: a method that boosts multimodal distillation for 3D semantic segmentation under domain shift.
📍 BMVC
🕚 Monday, Poster Session 1: Multimodal Learning (11:00–12:30)
📌 Hadfield Hall #859
We'll present MuDDoS at BMVC: a method that boosts multimodal distillation for 3D semantic segmentation under domain shift.
📍 BMVC
🕚 Monday, Poster Session 1: Multimodal Learning (11:00–12:30)
📌 Hadfield Hall #859
Reposted by Christian Wolf
We are introducing MuM, a feature encoder (ViT-L) tailored for 3D vision tasks.
TLDR; Spiritual successor to CroCo with a simpler multi-view objective and larger scale. Beats DINOv3 and CroCo v2 in RoMa, feedforward reconstruction, and rel. pose.
arxiv.org/abs/2511.17309
github.com/davnords/mum
TLDR; Spiritual successor to CroCo with a simpler multi-view objective and larger scale. Beats DINOv3 and CroCo v2 in RoMa, feedforward reconstruction, and rel. pose.
arxiv.org/abs/2511.17309
github.com/davnords/mum
November 24, 2025 at 10:27 AM
We are introducing MuM, a feature encoder (ViT-L) tailored for 3D vision tasks.
TLDR; Spiritual successor to CroCo with a simpler multi-view objective and larger scale. Beats DINOv3 and CroCo v2 in RoMa, feedforward reconstruction, and rel. pose.
arxiv.org/abs/2511.17309
github.com/davnords/mum
TLDR; Spiritual successor to CroCo with a simpler multi-view objective and larger scale. Beats DINOv3 and CroCo v2 in RoMa, feedforward reconstruction, and rel. pose.
arxiv.org/abs/2511.17309
github.com/davnords/mum
I have no idea how accurate this is, but for me it's a positive surprise. I had the impression that nothing had actually changed at all.
This, by @giacgrassi.bsky.social, remains one of the simplest and best climate visuals
November 23, 2025 at 10:18 PM
I have no idea how accurate this is, but for me it's a positive surprise. I had the impression that nothing had actually changed at all.
An interesting read.
November 23, 2025 at 9:59 PM
An interesting read.
TIL that "pico-bello" is actually not an Italian expression but only used in German 🇩🇪 (where we commonly use it to say perfect/clean/spotless). Its origins are from "piekfein" (= very fine), taking the first part of it, and adding the Italian 🇮🇹 "bello" to it.
Etymology is fascinating.
Etymology is fascinating.
November 23, 2025 at 4:27 PM
TIL that "pico-bello" is actually not an Italian expression but only used in German 🇩🇪 (where we commonly use it to say perfect/clean/spotless). Its origins are from "piekfein" (= very fine), taking the first part of it, and adding the Italian 🇮🇹 "bello" to it.
Etymology is fascinating.
Etymology is fascinating.
Reposted by Christian Wolf
"We thank the reviewers for their helpful suggestions"
November 23, 2025 at 11:19 AM
"We thank the reviewers for their helpful suggestions"
The future is here... Somewhat. Apparently you can set up a company with only AI employees talking to each other and actually doing stuff by querying the web and working on spreadsheets. Also,they share their trips to mountains they allegedly did during weekends. Hallucinations ensure.
Sam Altman says the one-person billion-dollar company is coming. Maybe I could be that person—if only I could get my colleagues to shut up and stop lying.
All of My Employees Are AI Agents, and So Are My Executives
Sam Altman says the one-person billion-dollar company is coming. Maybe I could be that person—if only I could get my colleagues to shut up and stop lying.
wrd.cm
November 22, 2025 at 11:03 PM
The future is here... Somewhat. Apparently you can set up a company with only AI employees talking to each other and actually doing stuff by querying the web and working on spreadsheets. Also,they share their trips to mountains they allegedly did during weekends. Hallucinations ensure.
Reposted by Christian Wolf
The 4th #AI4RoboticsWorkshop is over! A big THANK YOU to all our fab speakers & participants for great presentations & conversations. Snapshot souvenir ⬇️ For those who missed out -recordings available soon! tinyurl.com/bdtk2nzs
November 21, 2025 at 4:00 PM
The 4th #AI4RoboticsWorkshop is over! A big THANK YOU to all our fab speakers & participants for great presentations & conversations. Snapshot souvenir ⬇️ For those who missed out -recordings available soon! tinyurl.com/bdtk2nzs
Justin Carpentier on simulation, data, optimal control and RL, and on the fantastic Pinocchio simulator and optimization suite.
November 21, 2025 at 4:02 PM
Justin Carpentier on simulation, data, optimal control and RL, and on the fantastic Pinocchio simulator and optimization suite.
Nicolas Mansard: in locomotion / legged motion, when we worked on MPC (model predictive control) we were a team of 10 working on the robot and it barely worked. Now with RL (Reinforcement Learning) a single student can get a policy in 3 months, including training of the student.
November 21, 2025 at 2:56 PM
Nicolas Mansard: in locomotion / legged motion, when we worked on MPC (model predictive control) we were a team of 10 working on the robot and it barely worked. Now with RL (Reinforcement Learning) a single student can get a policy in 3 months, including training of the student.
Reposted by Christian Wolf
Changing a bicycle rotor with two manipulators working together as an ... unseen task (gasp) is crazy stuff.
Large Behavior Models (LBM) by TRI
Presented by Adrien
toyotaresearchinstitute.github.io/lbm1/
Large Behavior Models (LBM) by TRI
Presented by Adrien
toyotaresearchinstitute.github.io/lbm1/
November 21, 2025 at 2:15 PM
Changing a bicycle rotor with two manipulators working together as an ... unseen task (gasp) is crazy stuff.
Large Behavior Models (LBM) by TRI
Presented by Adrien
toyotaresearchinstitute.github.io/lbm1/
Large Behavior Models (LBM) by TRI
Presented by Adrien
toyotaresearchinstitute.github.io/lbm1/
Adrien Gaidon shares key messages for projects working on AI for Robotics:
1. Don't fool yourself (with evaluation).
2. Since you decided to rigorously evaluate, you can iterate quickly.
3. Engineering is still key, software is important.
4. Get the Data right
1. Don't fool yourself (with evaluation).
2. Since you decided to rigorously evaluate, you can iterate quickly.
3. Engineering is still key, software is important.
4. Get the Data right
November 21, 2025 at 2:06 PM
Adrien Gaidon shares key messages for projects working on AI for Robotics:
1. Don't fool yourself (with evaluation).
2. Since you decided to rigorously evaluate, you can iterate quickly.
3. Engineering is still key, software is important.
4. Get the Data right
1. Don't fool yourself (with evaluation).
2. Since you decided to rigorously evaluate, you can iterate quickly.
3. Engineering is still key, software is important.
4. Get the Data right
A speaker: "I am also active on social networks, at least the nice ones."
November 21, 2025 at 1:20 PM
A speaker: "I am also active on social networks, at least the nice ones."
Aniruddha Kembhavi presents Wayve's efforts on GAIA world models and their usage for not only deriving policies, but also evaluation. They can even perform in-context learning, videos of applying failure cases to new initial conditions.
AI 4 robotics workshop at @naverlabseurope.bsky.social
AI 4 robotics workshop at @naverlabseurope.bsky.social
November 21, 2025 at 1:18 PM
Aniruddha Kembhavi presents Wayve's efforts on GAIA world models and their usage for not only deriving policies, but also evaluation. They can even perform in-context learning, videos of applying failure cases to new initial conditions.
AI 4 robotics workshop at @naverlabseurope.bsky.social
AI 4 robotics workshop at @naverlabseurope.bsky.social
Eric Brachmann @ericbrachmann.bsky.social on the future of mapping ... or localization w/o explicit maps, and links to SfM.
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
November 21, 2025 at 12:46 PM
Eric Brachmann @ericbrachmann.bsky.social on the future of mapping ... or localization w/o explicit maps, and links to SfM.
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
Daniel Cremers presents the history of SLAM and his contributions, eg. the new VISTA SLAM (3DV 2026).
AI 4 Robotics workshop at @naverlabseurope.bsky.social
AI 4 Robotics workshop at @naverlabseurope.bsky.social
November 21, 2025 at 8:46 AM
Daniel Cremers presents the history of SLAM and his contributions, eg. the new VISTA SLAM (3DV 2026).
AI 4 Robotics workshop at @naverlabseurope.bsky.social
AI 4 Robotics workshop at @naverlabseurope.bsky.social
Fatma Guney on thinking fast and slow for autonomous driving: how to combine bigger models with higher latency with smaller models with lower latency by forecasting rich but stale features to the future.
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
November 20, 2025 at 2:40 PM
Fatma Guney on thinking fast and slow for autonomous driving: how to combine bigger models with higher latency with smaller models with lower latency by forecasting rich but stale features to the future.
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
Donghwan Lee presents the AI for robotics activities of Naver Labs Korea, here for example the smartphone based AR navigation application released to the general public recently, which is based on digital twins and visual localization.
AI 4 Robotics Workshop at @naverlabseurope.bsky.social.
AI 4 Robotics Workshop at @naverlabseurope.bsky.social.
November 20, 2025 at 1:53 PM
Donghwan Lee presents the AI for robotics activities of Naver Labs Korea, here for example the smartphone based AR navigation application released to the general public recently, which is based on digital twins and visual localization.
AI 4 Robotics Workshop at @naverlabseurope.bsky.social.
AI 4 Robotics Workshop at @naverlabseurope.bsky.social.
Cordelia Schmid on learning manipulation: combining 3D models, diffusion and LLMs, and also how to integrate a critic of the planning process.
AI 4 Robotics workshop at @naverlabseurope.bsky.social
AI 4 Robotics workshop at @naverlabseurope.bsky.social
November 20, 2025 at 11:48 AM
Cordelia Schmid on learning manipulation: combining 3D models, diffusion and LLMs, and also how to integrate a critic of the planning process.
AI 4 Robotics workshop at @naverlabseurope.bsky.social
AI 4 Robotics workshop at @naverlabseurope.bsky.social
Andrea Vedaldi talks about 4D reconstruction, leveraging 4D generation, but also representations based on dynamic point maps which encode points for different viewpoints AND different time steps.
AI 4 Robotics workshop at @naverlabseurope.bsky.social
AI 4 Robotics workshop at @naverlabseurope.bsky.social
November 20, 2025 at 11:21 AM
Andrea Vedaldi talks about 4D reconstruction, leveraging 4D generation, but also representations based on dynamic point maps which encode points for different viewpoints AND different time steps.
AI 4 Robotics workshop at @naverlabseurope.bsky.social
AI 4 Robotics workshop at @naverlabseurope.bsky.social
Reposted by Christian Wolf
We’re live! 🚀 Streaming: tinyurl.com/bdtk2nzs
The International Workshop on AI4Robotics by @naverlabseurope
2dys of Spatial AI, SLAM, robot learning, HRI, autonomy
This AM CET: @martinhumenberger.bsky.social @marcpollefeys.bsky.social Andrea Vedaldi Cordelia Schmid & @andrewdavidson.bsky.social ⬇️
The International Workshop on AI4Robotics by @naverlabseurope
2dys of Spatial AI, SLAM, robot learning, HRI, autonomy
This AM CET: @martinhumenberger.bsky.social @marcpollefeys.bsky.social Andrea Vedaldi Cordelia Schmid & @andrewdavidson.bsky.social ⬇️
November 20, 2025 at 8:40 AM
We’re live! 🚀 Streaming: tinyurl.com/bdtk2nzs
The International Workshop on AI4Robotics by @naverlabseurope
2dys of Spatial AI, SLAM, robot learning, HRI, autonomy
This AM CET: @martinhumenberger.bsky.social @marcpollefeys.bsky.social Andrea Vedaldi Cordelia Schmid & @andrewdavidson.bsky.social ⬇️
The International Workshop on AI4Robotics by @naverlabseurope
2dys of Spatial AI, SLAM, robot learning, HRI, autonomy
This AM CET: @martinhumenberger.bsky.social @marcpollefeys.bsky.social Andrea Vedaldi Cordelia Schmid & @andrewdavidson.bsky.social ⬇️
Marc Pollefeys talks about Global SfM meeting feedforward reconstruction, targeting scenes with a high number of hops between cameras to cover the entire scene (and many other Spatial AI contributions)
@marcpollefeys.bsky.social
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
@marcpollefeys.bsky.social
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
November 20, 2025 at 9:34 AM
Marc Pollefeys talks about Global SfM meeting feedforward reconstruction, targeting scenes with a high number of hops between cameras to cover the entire scene (and many other Spatial AI contributions)
@marcpollefeys.bsky.social
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
@marcpollefeys.bsky.social
AI 4 Robotics Workshop at @naverlabseurope.bsky.social
Andrew Davison presents his vision of the computational structures of Spatial AI.
AI 4 Robotics Workshop
AI 4 Robotics Workshop
November 20, 2025 at 8:58 AM
Andrew Davison presents his vision of the computational structures of Spatial AI.
AI 4 Robotics Workshop
AI 4 Robotics Workshop