About A collection of AWESOME things about information geometry Topics

Methods for computational information geometry

A Fisher information matrix python package for GW detector networks.

Lecture notes of Professor Stéphane Mallat - Collège de France - Paris

[ICML 2025] Official implementation of "Efficiently Access Diffusion Fisher: Within the Outer Product Span Space".

A framework for determining the maximum information gain and optimising experimental design in neutron reflectometry using the Fisher information.

Active learning for neural networks

[COLM 2025] Improving Fisher Information Estimation and Efficiency for LoRA-based LLM Unlearning

Experimental design in neutron reflectometry using the Fisher information.

Minimal effort to get maximal information from your experiments

My PhD thesis: Regression modelling using priors depending on Fisher information covariance kernels (I-priors)

Replication package for Fisher curvature scaling exponent d_R = 10/9 at 2D Ising critical point

Verification harness for quantum ML. A reproducible lab for stress-testing quantum models where predictive accuracy, identifiability, curvature, and robustness under noise can diverge.

Matlab package for LNA simulations of biological models

Closed-form MLE and Fisher information for exponential series systems with masked failure data. Includes theoretical results, proofs, and numerical validation.

An R package for I-prior regression

A sensitivity toolbox that is tailored to the design process in the presence of uncertainties

A hands‑on, first‑principles guide to fitting logistic regression via the Iteratively Reweighted Least Squares (IRLS) algorithm complete with mathematical derivations, R code from scratch, and a real‑world S&P data case study to bring your statistical modeling skills to the next level.

Cadre théorique formel visant à unifier la thermodynamique, la théorie de l’information et la gravitation émergente à travers la géométrie informationnelle. Le projet explore la courbure thermodynamique de Ruppeiner ($R$) comme invariant géométrique encodant la microstructure, les corrélations et les interactions fondamentales des systèmes physique

Graph coherence engine for research datasets — Fisher Information diagnostics, two-tier reports, claim extraction

No description provided.

Data and code modeling cell signal processing using maximum-likelihood estimators, information theory, and stochastic differential equation simulations for "Physical limits of galvanotaxis"

Quantum computer simulator with information geometry interpretation. Features Lindblad dynamics, spectral zeta functions, and geodesic-spectral duality connecting decoherence to operator spectra.

Résolution du Paradoxe de Levinthal et de l'Abiogenèse par la Frustration Géométrique. La vie comme état métastable de la matière, pilotée par la minimisation de l'énergie libre sur des paysages rugueux.

PyTorch implementation of the paper "FISH-Tuning: Enhancing PEFT Methods with Fisher Information" (2025). A drop-in improvement for LoRA that automatically freezes unimportant weights. Helps you save memory and train highly sparse models without losing accuracy.

Framework de physique théorique unifiant Relativité, Mécanique Quantique et Thermodynamique via la Géométrie de l'Information (Métrique de Fisher/Ruppeiner). Une approche moniste où la réalité émerge de la compression informationnelle (κ).

Additive interaction modelling using I-priors

Unified theoretical–empirical verification of the Cognitive Uncertainty Principle (CUP). A jump–diffusion model reveals an epistemic phase boundary between a Fisher regime (Δε·ΔDₖₗ ≥ 1.17×10⁻⁴) and a KL regime (Δε·ΔDₖₗ ≈ 1.71×10⁻²). The canonical bound Δβ·KL ≥ 3.94×10⁻⁴ remains robust.

Reproducible data processing and analysis pipeline for the manuscript "Superconducting antiqubit simulation enables optimal phase estimation via unitary inversion".

KL-Geometric Structure of Observer Entropy. Bridge Theorem (Khomyakov's Bridge): S_obs = ½ε²vᵀI(θ)v + O(ε³). Fisher–Rao metric, sufficient conditions, dissipation functional. Two worked examples (4-point and 5-point spaces). Python verification script and figures.