Introduction to Predictive Coding Networks for Machine Learning
This is an incremental introduction for machine learning practitioners interested in biologically inspired frameworks.
The paper introduces predictive coding networks (PCNs) as a biologically inspired alternative to feedforward neural networks, providing a quick introduction for practitioners with a benchmark-smashing application on CIFAR-10 image classification.
Predictive coding networks (PCNs) constitute a biologically inspired framework for understanding hierarchical computation in the brain, and offer an alternative to traditional feedforward neural networks in ML. This note serves as a quick, onboarding introduction to PCNs for machine learning practitioners. We cover the foundational network architecture, inference and learning update rules, and algorithmic implementation. A concrete image-classification task (CIFAR-10) is provided as a benchmark-smashing application, together with an accompanying Python notebook containing the PyTorch implementation.