About A collection of AWESOME things about information geometry Topics

Code accompanying our papers on the "Generative Distributional Control" framework

Official implementation of the NeurIPS 24 paper of statistical flow matching (SFM) for discrete generation.

Official code for Fisher information embedding for node and graph learning (ICML 2023)

Embedding language models in probability space via log-likelihood vectors

Research-grade PyTorch math: differential geometry, spectral graph theory, discrete Ricci flow, simplicial topology, persistent homology, cellular sheaves, SO(3) Lie primitives, information geometry, tensor decompositions, content-addressable provenance. GPU-native, batched-first, audit-clean, cited.

[ICLR 2026] Official implementation of the paper: Learning a distance measure from the information-estimation geometry of data

Fisher-Bures Adversary Graph Convolutional Networks

PyTorch implementation of α-geodesical skew divergence

This is an awesome list for information geometry. This is my curation of resources on this topic as such there are bound to be things I missed, please contribute if you think I missed anything.

Picture of the Space of Learnable Tasks (ICML 23)

Essential Books for Computer Science

this project is a simulation of Signal Detection for Ultra-Massive MIMO: An Information Geometry Approach paper. This is my project for Detection and Estimation class.

Lean 4 library characterizing the algebraic structure of metastability and consolidation in stochastic systems

A barrier-coordinate theory of existential risk measurement.

This repository contains coursework and project materials for an advanced course on Information Geometry, Statistics and Learning. It includes problem set solutions and a final project exploring Information geometry and it's applications. The work bridges differential geometry with statistics and machine learning.

The Golomb Universe

Decoder-only Transformer where attention operates on geodesic distances in a learned Riemannian manifold with gravitational curvature and variable dimensionality per token. Based on Directional Relational Manifolds (DRM)

Lean verified Science.

Curriculum learning framework that uses geometrically structured datasets and lets the model self discover learning path through emergent dynamics