Practical Spiking Neural Networks

A hands-on introduction to Spiking Neural Networks (SNNs) that teaches you to design, build, train, and deploy neuromorphic systems - both in simulation and on hardware.

What is this book about?¶

SNNs are biologically-inspired neural networks that use discrete spikes for computation, making them inherently sparser and more energy-efficient than conventional ANNs when deployed on neuromorphic hardware. Computationally, SNNs are also more expressive because they combine numerical calculus (in the continuous domain $\mathbb{R}$) with discrete logic (in the discrete domain $\mathbb{N}$) Neumann, 2012.

This first of its kind, community-driven and open-source book on SNNs bridges theory and practice with interactive examples, guiding you from SNN fundamentals through training and optimization methods to neuromorphic hardware deployments. We cover three topics:

1. Fundamentals: Covered in the Topic 1: Foundations of SNNs
2. Training and Optimization: Covered in the Topic 2: Training SNNs
3. Neuromorphic Hardware: Covered in the Topic 3: Deploying SNNs

How is this book written?¶

This book is completely open-source and is freely available for everyone. It is currently being actively written by a group of nearly 10 researchers, who range from late stage PhD candidates to scientists with decades of neuromorphic experience. The contents of this book will be reviewable by all, thus, vetted by the wider neuromorphic community for correctness.

Contributing¶

The book is open source under CC BY 4.0. Once the version 1.0 of the book is released, we will welcome improvements and pull requests. If you are interested to contribute by writing or reviewing, read more in the contribution guidelines and reach out to us on the Open Neuromorphic Discord server.

Citing¶

This book will be made citable as a whole and as individual chapters. We very much encourage citing specific chapters to credit the contributors who helped your research. Upon the release of version 1.0, the bibtex of the book and individual chapters will be made available.

Editors¶

• Ramashish Gaurav
• Jens Egholm Pedersen
• Petrut A. Bogdan