Symplectic Reservoir Representation of Legendre Dynamics Modern learning systems act on internal representations of data, yet how these representations encode underlying physical or statistical structure is often left implicit. In physics, conservation laws...

Beyond metric preservation: build symplectic structural invariance into representation.

arxiv.org/abs/2512.19409

#ReservoirComputing #RepresentationLearning #InformationGeometry #SymplecticGeometry #HamiltonianDynamics #GeometricDeepLearning #DynamicalSystems #PhysicsInformedML

De la dispersion des flux à l'Equivalence : E = m*c² de A.Einstein

💡via E.G Fischer (1813):
Boulliau: I ∝ 1/d²
Rømer: v = d/t
Galilée: S = ½gT², V = √(2gS)
Kepler-Rømer: L ∝ 1/d²
Newton: F = m*a
Lambert-Bouguer: L:l = (AB²):(AE²)

Boulliau → Newton

#InformationGeometry #propagation #diffusion

Fisher Co‑metric: New Perspective on Cotangent Bundles

The Fisher co‑metric, defined directly on the cotangent bundle, links variance to geometry; the revised paper was posted on 4 Oct 2025 and appears in Information Geometry (2024). Read more: getnews.me/fisher-co-metric-new-per... #informationgeometry #fishermetric

Information Geometry and Variational Bayes: Natural Gradients for AI

Researchers show variational Bayes often needs natural‑gradient calculations; the Bayesian Learning Rule (BLR) from 2023 enables inference for models with billions of parameters. getnews.me/information-geometry-and... #informationgeometry #bayes

Information structures and their cohomology We introduce the category of information structures, whose objects are suitable diagrams of measurable sets that encode the possible outputs of a given family of observables and their mutual relations...

Wanting to geek out on information geometry I went looking for ways to understand this through category theory.

Information Structure and their cohomology by Juan Pablo Vigneaux.

Pretty dense, working on a visualization.

arxiv.org/abs/1709.07807

#informationgeometry #categorytheory #math #AI

I'm looking to gather material about the practical importance of the sign of curvatures in #informationgeometry .. what are the consequences in practical statistical estimation, if we know that the curvature in the regions of interest is positive/negative? Anyone knows about this?

Mystical World of Information Geometry Tales of geometry for information theory and machine learning

Mystical World of Information Geometry.
Information geometry is a mathematical tool for exploring the world of information using geometry.
towardsdatascience.com/mystical-wor...
#informationgeometry #geometry #informationtheory #datascience #machinelearning #ml #Ai #optimization #topology

Original post on mastodon.social

Who am I, and why am I here? #introduction

I am a machine learning researcher, using tools from #Bayes, #stats, #optimization, #informationgeometry, #deeplearning, signal processing, etc.

I care deeply about people, their well-being, inclusion, diversity, equity, privacy, and justice.

I […]