This repository serves as a curated, up-to-date list of research papers, surveys, tools, and tutorials focused on hyperbolic representation learning and deep learning. It aims to assist researchers and practitioners in navigating the rapidly evolving field of hyperbolic geometry for machine learning, particularly for data exhibiting hierarchical or power-law structures.

How It Works

The project meticulously categorizes a vast collection of academic literature related to hyperbolic spaces. It covers foundational concepts, various hyperbolic models (shallow, neural networks, GNNs, Transformers), and their applications across diverse domains like NLP, computer vision, and recommender systems. The organization facilitates easy discovery of relevant research and promotes community engagement.

Quick Start & Requirements

This is a curated list of papers, not a software library. No installation or execution is required.

Highlighted Details

• Comprehensive categorization of papers from shallow models to LLMs in hyperbolic space.
• Includes surveys, books, tools (e.g., Geoopt, GraphZoo), and tutorials.
• Regularly updated with papers from major ML conferences (ICLR, NeurIPS, ICML, etc.).
• Covers a wide array of applications including NLP, Computer Vision, Recommender Systems, and Knowledge Graphs.

Maintenance & Community

The repository is actively maintained, with recent updates including ICLR 2025 and NeurIPS 2024 papers. A Slack community is available for discussion and suggestions.

Licensing & Compatibility

The repository itself is not software and does not have a license. The linked papers are subject to their respective publication licenses.

Limitations & Caveats

As a curated list, it does not provide implementations or code. The breadth of the topic means some oversights or mistakes are possible, as noted by the maintainers.