I'm a bot that surveys the literature in geometric statistics and geometric (deep) learning! Operated by the Geomstats team, adapted from @fxcoudert's bot.
Joined August 2020
- Tweets488
- Following335
- Followers1.1K
- Likes537
Pinned Tweet
Our Python module "Information Geometry" is available!
With:
🌐The Fisher-Rao Riemannian manifolds of probability distributions such as: normal, Gamma, Dirichlet, and more
🌐The Fisher-Rao geometry of any parametric family of distributions of interest.
☀️New paper alert: “Parametric information geometry with the package Geomstats”
Work done with Alice Le Brigant, @jujuzor, and @ninamiolane.
ACM Transactions on Mathematical Software: dl.acm.org/doi/10.1145/36275…
Github: github.com/geomstats/geomsta…
geomstats retweeted
🗓️ Join me on October 7 for one of the inaugural lectures of the Science of Intelligence Institute at @UCBerkeley !
Can we write scientific laws for intelligence? I believe that we can, that the math for it already exists, and that they apply across brains & machines.
tinyurl.com/mvrfs39w
geomstats retweeted
Exciting news, geometers! 🌐
The Workshop Bridging Optimal Transport, Learning and Structured Data: Toward Geometric Distributional Learning, will take place in person on December 12 or 13 in Paris, France, as part of #NeurIPS 2026.
W/ @JustinMSolomon and team!
geomstats retweeted
Our review on superposition in AI & brains is published in @Nature #Machine #Intelligence 🌐
Neural networks encode concepts in superposition. We can recover them in 3 steps: identifiability, compressed sensing, interpretability.
W/ @klindt_david @oneill_c @rpatrik96 H Maurer🌟
Our review on superposition in AIs and brains is finally published in Nature Machine Intelligence 🙌
nature.com/articles/s42256-0…
geomstats retweeted
It decomposes any lifted dataset into its graph representation, containing all the information about rank (hyperedges? faces? volumes?) and neighborhoods (adjacencies? incidences?) that fully define that lifting. It's the map through which information will flow during training.
geomstats retweeted
Ali-Silvey-Csiszar f-divergence
I_f(p:q)=∫ p f(q/p) dμ
include Kullback-Leibler, Jeffreys, Jensen-Shannon, etc.
f-divergence as a Taylor series of *higher-order chi divergences*
Sufficient conditions for convergence of the series given in
👉tinyurl.com/fdivTaylorChi
geomstats retweeted
When neural networks are trained to compose a sequence of group elements, what features do they learn? The irreducible representations of the group!
Very happy that our work led by @giovannimarchet and @KuninDaniel has been accepted to @icmlconf
Beautiful application of group theory to learning mechanics!
also with @AdeleMyersPhD and @ninamiolane
Excited to share that our paper “Sequential Group Composition: A Window into the Mechanics of Deep Learning” was accepted to ICML 2026 in Seoul!
Co-led with @giovannimarchet
and @AdeleMyersPhD @hopfbifurcator @ninamiolane
Paper: arxiv.org/abs/2602.03655
geomstats retweeted
Using the AGF framework (arxiv.org/abs/2506.06489) we developed in prior work, we show how two-layer networks learn this task one irreducible representation at a time.
The order in which these features are learned is determined by the Fourier statistics of the encoding
geomstats retweeted
Excited to share that our paper “Sequential Group Composition: A Window into the Mechanics of Deep Learning” was accepted to ICML 2026 in Seoul!
Co-led with @giovannimarchet
and @AdeleMyersPhD @hopfbifurcator @ninamiolane
Paper: arxiv.org/abs/2602.03655
geomstats retweeted
Arithmetic. Vision. Navigation. Planning. What if they're all the same task?
We introduce group composition as a unifying abstraction for many learning problems & show neural networks crack it using Fourier!
Led by @KuninDaniel @giovannimarchet @AdeleMyersPhD @hopfbifurcator🌟
Excited to share that our paper “Sequential Group Composition: A Window into the Mechanics of Deep Learning” was accepted to ICML 2026 in Seoul!
Co-led with @giovannimarchet
and @AdeleMyersPhD @hopfbifurcator @ninamiolane
Paper: arxiv.org/abs/2602.03655
geomstats retweeted
🚨One of the coolest workshops in AI is back!
NeurReps 2025 is calling for papers on symmetry, geometry & topology in neural networks 🧠📐
If your work bridges theory & structure — don’t miss this.
📅 Deadline: Aug 22
📢 Call for Papers: NeurReps 2025 ‼️‼️‼️
🧠 Submit your research on symmetry, geometry, and topology in artificial and biological neural networks.
Two tracks: Proceedings (9 pages) and Extended Abstract (4 pages).
Deadline: Aug 22, 2025.
neurreps.org/call-for-papers
geomstats retweeted
Want to learn more?🧐
📺 Subscribe to the @neur_reps YouTube channel and find more talks by @mweber_PU @KostasPenn @RobinSFWalters @erikjbekkers S. Ravanbakhsh @t_andy_keller & more!
youtube.com/@neurreps
🌐 Follow our lab @geometric_intel @ai_ucsb :
gi.ece.ucsb.edu/
geomstats retweeted
🤖 New Talk: Geometric Intelligence in Minds & Machines🧠 Available on Youtube:
youtube.com/watch?v=1RtHB24s…
@louisacornelis & I opened the 2025 @neur_reps Global Speaker Series @geometric_intel @ai_ucsb @UCSBengineering
Thanks @mathildepapillo @ChasevdG @BTolooshams for hosting!
📢 NeurReps Speaker Series seminar is back!
@UCSBEngineering is hosting us for a talk by @ninamiolane and @louisacornelis on Geometric Intelligence in Minds and Machines!🧠
🗓️This Friday, February 21st, 11 AM PT
Join in person at @UCSB (Elings, 1605) or on Zoom!
geomstats retweeted
Thank you to everyone who showed up in person and online! Congrats to @ninamiolane and @louisacornelis for some stellar talks🤩
I am excited to host the next edition of the NeurReps Global Seminar Series, from @neur_reps !
@ninamiolane and @louisacornelis will be speaking on 🧠Geometric Intelligence in Minds and Machines 🤖
📅Friday, Feb 21st, 11am PST
Join us at UCSB (Elings, 1605) or on Zoom!
geomstats retweeted
Can't believe it's a year already! Incredibly proud of what this team has accomplished 🧠🌟
@emilyjacobs @caitaylo @russpoldrack @amykooz @crntozlu @joshbuck @kbcasaletto @bucklr01 @PaLab_UCSD Suzanne Baker @susanna_carmona @MagdaMartinezGa @AdeleMyersPhD @louisacornelis
geomstats retweeted
Hi everyone - I'm on bsky🦋
Hope to see you all there! Same handle🌟
bsky.app/profile/ninamiolane…
geomstats retweeted
We’re inspired + energized after three days of roundtable discussions, keynotes & lightning #tech talks about the challenges and future of imaging.
#ImagingTheFuture
geomstats retweeted
🏆Geometric AI for weather forecasting is awarded TIME Invention of the year!🌎
Incredibly proud to serve as advisor for @atmo_ai - the leading startup in AI & Weather.
Congrats @johmathe @alevy @YvesMartinDT @hopfbifurcator & team!
*This* is AI for science 🌟
Atmo is one of the TIME best inventions of 2024!
We are on a mission to make ultra-precise weather forecasting universally accessible. Atmo is honored to be recognized by TIME for our contribution.
See the full list here: time.com/best-inventions-202…
geomstats retweeted
The Philippines just received the first live AI-based Weather Forecasting System in the region.
atmo.ai/news/philippines-ai-…
As extreme weather events become more frequent, AI weather forecasting system become critical parts of early warning systems (EWS).
geomstats retweeted
Atmo is one of the TIME best inventions of 2024!
We are on a mission to make ultra-precise weather forecasting universally accessible. Atmo is honored to be recognized by TIME for our contribution.
See the full list here: time.com/best-inventions-202…
geomstats retweeted
Very proud of the Atmo team and the recognition from TIME. time.com/collection/best-inv…
Atmo is one of the TIME best inventions of 2024!
We are on a mission to make ultra-precise weather forecasting universally accessible. Atmo is honored to be recognized by TIME for our contribution.
See the full list here: time.com/best-inventions-202…