This repository is a curated reading list for state-space models (SSMs) in machine learning, targeting researchers and practitioners interested in this alternative to Transformers for sequence modeling. It provides a comprehensive overview of foundational concepts, architectures, and applications across various domains like language, vision, audio, and time-series.

How It Works

The list categorizes resources by topic, including tutorials, surveys, books, and specific model architectures like Mamba and S4. It highlights key papers and code implementations, offering a structured path to understanding the evolution and capabilities of SSMs, particularly their efficiency in handling long sequences compared to traditional attention mechanisms.

Quick Start & Requirements

This is a curated list of resources, not a runnable codebase. No installation or specific requirements are needed to browse the content. Links to papers, code repositories, and blog posts are provided.

Highlighted Details

• Extensive coverage of Mamba and S4 architectures, including theoretical explanations and implementation guides.
• Categorization of resources across diverse ML domains: Vision, Language, Audio, Time-Series, Medical, Tabular, and Reinforcement Learning.
• Inclusion of both foundational classical state-space models and modern deep learning approaches.
• Links to code repositories for many featured papers, enabling practical exploration.

Maintenance & Community

Contributions are welcome via pull requests following contribution guidelines. The repository is actively maintained, with community contributions encouraged for expanding the list.

Licensing & Compatibility

The repository itself is licensed under the MIT License. Individual resources linked within the list are subject to their respective licenses.

Limitations & Caveats

This is a reading list and does not provide a unified framework or codebase for experimenting with state-space models. Users must individually locate and set up the code for each paper of interest.