Giovanni Beltrame

Took some time, but we have released the code for our swarm distributed caging and transport system. Info and paper on GitHub:
github.com/MISTLab/Swar...

🧠💤 New research in personalized #BrainStimulation using the Portiloop device: our latest paper shows how #AI can learn to improve detection of #sleep spindles (EEG signals linked to memory consolidation) for each specific subject.
📖 Paper: iopscience.iop.org/article/10.1...
#PersonalizedMedicine

Professor - Canada Excellence Research Chair for Critical Innovations to Strengthen Canadian Technology Sovereignty Polytechnique Montréal is an internationally recognized engineering university. Located in the heart of Montréal on Mount Royal, it is renowned for the high quality of the training offered at all levels, and for its multidisciplinary and multisectoral research, at the forefront of the local,...

An excellent opportunity for world-calibre scientists. Come and join us @polymtl.bsky.social !

www.polymtl.ca/carriere/en/...

3/3 We also provide ad-hoc networking and multi-robot coordination for mission planning. All code is #opensource and available for download on GitHub.

Project page: mistlab.ca/MGGPlanner/
Paper: ieeexplore.ieee.org/iel8/7083369...
Code: github.com/MISTLab/MGGP...
Video: youtu.be/Fv8B0Ml0KCY?...

2/3 We tested the planner at the Canadian Space Agency Mars Yard with multiple rovers using lidar-inertial sensing and show superior exploration coverage as well as 50% lower runtime compared to state-of-the-art methods.

1/3 We have released our #multi-robot #exploration #planner for #space applications! Designed for complex unconstrained environments, it mixes a local grid with a global graph planner, with the global graphs being synchronised when robots are in comms range. #field #robotics #autonomous #systems

Pls share: Konstanz School of Collective behavior from 21st July - 8th August. Application deadline: April, 15. An amazing line up of speakers spanning all aspects of collective behavior . For info: www.exc.uni-konstanz.de/kscb/
@cbehav.bsky.social @mpi-animalbehav.bsky.social

Giovanni Beltrame (Polytechnique Montreal), "Field collective robotics: Challenges and applications" - scienceofintelligence.de Giovanni Beltrame is a Professor in the Department of Computer and Software Engineering at Polytechnique Montréal, where he leads the Making Innovative Space

Giving a talk at the #Science of #Intelligence cluster at TU Berlin and online (Zoom) next week, "Field Collective Robotics: Challenges and Applications". You're welcome to join to see some of our recent developments.

www.scienceofintelligence.de/event/giovan...

VLMs everywhere, I'm curious to see where these models for navigation will lead. Interesting work from NVIDIA/UCSD:

Website: navila-bot.github.io
arXiv: arxiv.org/abs/2412.04453

A global horizon scan of the future impacts of robotics and autonomous systems on urban ecosystems - Nature Ecology & Evolution The future challenges and potential opportunities of robotics and autonomous systems in urban ecosystems, and how they may impact biodiversity, are explored and prioritized via a global horizon scan o...

The rise of robotics 🤖 and autonomous systems is transforming cities 🌆! This global study identifies 32 opportunities, such as advanced eco-monitoring 🌿, and 38 challenges, including RAS waste 🛠️. It highlights the need for proactive policies to balance urban innovation with sustainability 🌍.
#robot

A caving expedition yields valuable insight into the challenges of field research - Concordia University Learning how to practice science in extreme environments offers lessons far beyond the lab, new Concordia study shows

Our recent cave neuroscience expedition, with @hrjourde.bsky.social and @beltrame.mistlab.ca:

www.concordia.ca/news/stories...

See also our thoughts about how to take cognitive neuroscience outside of the box!
www.sciencedirect.com/science/arti...

#cognition #speleology #interdisciplinary

𝗦𝗲𝗹𝗳-𝗢𝗿𝗴𝗮𝗻𝗶𝘇𝗶𝗻𝗴 𝗦𝘆𝘀𝘁𝗲𝗺𝘀: 𝗪𝗵𝗮𝘁, 𝗛𝗼𝘄, 𝗮𝗻𝗱 𝗪𝗵𝘆?
Good intro by Carlos Gershenson
#complexity
www.preprints.org/manuscript/2...

I'm really excited about structure from motion, which is having a real renaissance (ACE0, FlowMap, MASt3R-SfM, VGGSfM, GLOMAP, etc). It will be cool to see how SfM methods interact with VLMs -- SfM always seems like the last problem to jump on the latest bandwagon, but maybe that will change.

Hello and welcome! (I'm one of your colleagues :)

To quote Douglas Adams, I love the whooshing sound deadlines make as they fly by

Beyond A*: Better Planning with Transformers via Search Dynamics Bootstrapping While Transformers have enabled tremendous progress in various application settings, such architectures still trail behind traditional symbolic planners for solving complex decision making tasks. In t...

The Beyond A* paper is fascinating. In it they use algorithmic techniques (A*) to solve a problem to produce a large dataset of training data to use to train a neural network to reach the same level of performance as the algorithm… 1/2

arxiv.org/abs/2402.14083

Thank you to the @neuripsconf.bsky.social for this recognition of the Generative Adversarial Nets paper published ten years ago with @ian-goodfellow.bsky.social, Jean Pouget-Abadie, @memimo.bsky.social, Bing Xu, David Warde-Farley, Sherjil Ozair and Aaron Courville.
blog.neurips.cc/2024/11/27/a...

The migration is happening!

Very interesting paper, not only for neuroscientists

🤔 Can you turn your vision-language model from a great zero-shot model into a great-at-any-shot generalist?

Turns out you can, and here is how: arxiv.org/abs/2411.15099

Really excited to this work on multimodal pretraining for my first bluesky entry!

🧵 A short and hopefully informative thread:

Voici un résumé des activités de commémoration du 6 décembre 1989 qui auront lieu au cours des prochains jours.
polymtl.info/Fp1bi

YouTube video by BMVA: British Machine Vision Association 1 Andrew Davison, Imperial College London - BMVA Symposium: Robotics Foundation & World Models

For my first post on Bluesky, this recent talk I did at the recent BMVA one day meeting on World Models is a good summary of my work on Computer Vision, Robotics and SLAM, and my thoughts on a bigger picture of #SpatialAI.
youtu.be/NLnPG95vNhQ?...

Introducing “MAGiC-SLAM: Multi-Agent Gaussian Globally Consistent SLAM”! We do SLAM with novel view synthesis capabilities on multiple simultaneously operating agents!

vladimiryugay.github.io/magic_slam/i...
1/7

My deep learning course at the University of Geneva is available on-line. 1000+ slides, ~20h of screen-casts. Full of examples in PyTorch.

fleuret.org/dlc/

And my "Little Book of Deep Learning" is available as a phone-formatted pdf (nearing 700k downloads!)

fleuret.org/lbdl/

Reinforcement Learning with Random Delays Action and observation delays commonly occur in many Reinforcement Learning applications, such as remote control scenarios. We study the anatomy of randomly delayed environments, and show that...

That's a great point! There has been a few works in this area (including ours) and there's a lot more to do:
Modeling RL with delays to maintain the Markov assumption
openreview.net/forum?id=QFY...
Learning variable control rates and adapting action duration
arxiv.org/html/2407.00...

Frustrating, yet incredibly exciting. The amount of progress in manipulation is fantastic. I wonder if they had any ripped shirts or things flying off the tables (probably).
Source of the video: www.physicalintelligence.company/blog/pi0

Check out the Distributed Robotics Systems Starter Pack:

It features researchers exploring swarm robotics, multi-robot systems, modular robots, and more!

Know someone who should be included? Suggest yourself or others in the replies! 🐝🤖

#Robotics #SwarmRobotics #DistributedRobotics #Swarm

That's an interesting idea! I think it falls under the umbrella of robotics for #climate #change: precision agriculture, recuperation of rare earths from tailing mounds, maintenance of clean energy infrastructure... Good to see people looking into this!

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.

The 2nd workshop on #Cooperative #Multi-Agent Systems #Decision-Making and #Learning: Human-Multi-Agent Cognitive Fusion will be held on 03/03/2025 in conjunction with #AAAI. We invite everyone to share your works as a paper contribution or an extended abstract.
Site: www.is3rlab.org/aaai25-cmasd...