This repository serves as a curated bibliography for research at the intersection of Category Theory and Machine Learning. It aims to consolidate papers exploring how abstract algebraic structures can provide foundational insights and novel approaches to ML problems, targeting researchers and practitioners seeking to leverage categorical frameworks for deeper understanding and more principled ML design.

How It Works

This repository is a collection of research papers, organized by sub-fields within machine learning where category theory is applied. The papers themselves explore how categorical concepts like functors, monads, spans, and sheaves can model ML components such as neural network architectures, learning dynamics, automatic differentiation, and probabilistic reasoning, offering a unified and compositional perspective.

Quick Start & Requirements

This repository is a list of papers and does not have a direct installation or execution process. Users are expected to access and read the linked research papers independently.

Highlighted Details

• Covers a broad spectrum of ML sub-fields including Deep Learning, Equivariance, Graph Neural Networks, Differentiable Programming, and Probabilistic Modeling.
• Features papers on foundational aspects like categorical deep learning, algebraic dynamical systems, and category theory for quantum natural language processing.
• Includes work on specific ML techniques such as attention mechanisms, neural operators, and automatic differentiation using categorical tools.
• Organized into thematic sections, facilitating targeted exploration of specific research areas.

Maintenance & Community

The repository is community-driven, encouraging contributions via pull requests or issues for missing papers or suggested changes. Specific contributors, maintainers, or community links (e.g., Discord, Slack) are