Steeven Janny

Research Scientist @NaverLabs Europe

[NAVER LABS Europe @naverlabseurope]

Incorporating physics into embodied AI has massive impact on out-of-the-box robot navigation! @jannysteeven & @chriswolfvision share what was learned in moving from simulation to the real world with #spatialAI ➡️https://t.co/4UCohg9J8I

[Christian Wolf (🦋🦋🦋) @chriswolfvision]

We have an open internship position on socially aware navigation, human aware world models etc.

At @naverlabseurope in Grenoble (Meylan), France.

https://t.co/W2sACGBzUH

[Christian Wolf (🦋🦋🦋) @chriswolfvision]

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.

https://t.co/lhI2qPySWH

Work by @JannySteeven et al.

[Steeven Janny @JannySteeven]

@3ar0bi I'm still using PowerPoint or Inkscape, but it's not great. I tried D3 also for that, but it's messy (it's basically boils down to write SVG)

[Steeven Janny @JannySteeven]

@3ar0bi @chriswolfvision Mostly caffeine and chatGPT...
We use d3js for most of the tools (and Blender for videos), none of us is really an expert in JS, so typically we kick start the code with LLM's and tweak the code until it works !

[Christian Wolf (🦋🦋🦋) @chriswolfvision]

Our e2d trained navigation agent has been accepted to #CVPR2025: fast-moving, efficient: correct modelling of motion during training in simulation.

https://t.co/lhI2qPySWH

By S Janny, H Poirier, L Antsfeld, G Bono, G Monaci, B Chidlovskii, F Giuliari, A Del Bue and myself.

[Steeven Janny @JannySteeven]

PhD Internship @naverlabseurope !

Our Spatial AI team is looking for a PhD intern to work on robot navigation! If you have experience in end-to-end training & RL, and want to work with real robots, DM me!

Bonus: great team, good coffee, amazing landscapes and french castle!

[Christian Wolf (🦋🦋🦋) @chriswolfvision]

At @naverlabseurope in Grenoble, France, we are searching for talented PhD interns for work on Spatial AI, geometric and robotic foundation models for navigation and manipulation. If you have experience in Embodied AI and are interested, DM me.

[Christian Wolf (🦋🦋🦋) @chriswolfvision]

Led by Guillaume Bono (@_WGW101), we have just released the code and weights for our SOTA image goal nav agent, published at #ICLR2024.

TLDR: take geometric foundation model, finetune it on rel pose + visbility, add to agent and adapt it with RL.

https://t.co/H3vysuFhwD

1/6

[Steeven Janny @JannySteeven]

Sometimes, you can't see it all...
In our #ICLR2024 spotlight paper, we show that you can still train a dense fluid simulator🌊 from super sparse measurements.

The trick was to rely on Control Theory to design the model!
Project page: https://t.co/VEQ6Ezr84r

[Christian Wolf (🦋🦋🦋) @chriswolfvision]

The PhD defense of Assem Sadek (@assemasadek) is happening right now, on "Building Autonomous Agents with Hybrid Navigation Policies". Committee: D. Filliat,
@FabienMoutarde, @BabelMarie, @ati_baskurt, B. Chidlovskii and myself.

@naverlabseurope @LIRISLyon @insadelyon

[Steeven Janny @JannySteeven]

@ProfJasonCorso @chriswolfvision Glad you liked it ! The slides are publicly available on my personal page (https://t.co/6HF9YUoBIe), as well as the sources. I hope you'll find a way to use them for your own purpose

[Steeven Janny @JannySteeven]

Oh my... Aren't you pretty ?! ♥️

[Steeven Janny @JannySteeven]

I made a funny viz tool using the ArXiv API, it queries ArXiv by keywords, and displays the authors sorted by number of papers. There is also a search form to quickly find people in the plot.
We found a reviewer for my Ph.D with it!
https://t.co/Iac4V1AUAv

[Steeven Janny @JannySteeven]

I realized that I am treating my bibliography like a wine cellar: I avoid the popular picks, like rare gems from smaller group with limited outputs and I am proud of my old forgotten vintages ! 🍷📚

[Steeven Janny @JannySteeven]

Playing around with fluid dynamics for a figure in the PhD manuscript. Same setup, but 4 different models of turbulences !

The difference between Large Eddy Simulation (LES) and RANS-based is impressive ! But the computation time is terrible...

[Steeven Janny @JannySteeven]

I'm pulling out Blender to illustrate some parts of the video presentation for #iclr2023 ! Well, this is not the most appetizing coffee ever... But go check our project page, simulations are WAY better looking !

https://t.co/YYXldfzwW5

[Christian Wolf (🦋🦋🦋) @chriswolfvision]

My colleagues M. Nadri, @PDufourControland L. Bako have a 6 month internship open on "Model predictive control of nonlinear systems based on affine differential inclusions", cont to PhD possible.

Lyon, France, https://t.co/BVfpCeHtBh
Contact: [email protected]

[Steeven Janny @JannySteeven]

@francoisfleuret You may use full names like W_\text{query}, that is heavier, but also easier to read...

[Steeven Janny @JannySteeven]

a BIG thank you to my co-authors: Aurélien Bénéteau, Madiha Nadri, @JulieDigne, @thomenicolas1 and @chriswolfvision !!

[Steeven Janny @JannySteeven]

Excited to present at #iclr2023 ! Our new fluid mechanics dataset, EAGLE, is a game changer with over 1.1 million super complex mesh simulations, streamlining future deep fluid simulators 🌊!

Project : https://t.co/Bm84d3nPQT
Paper : https://t.co/yLUbO0ixbt