Neural Network Basics

Status: public · Confidence: medium (0.84) · Basis: verified_sources

## TL;DR

Neural networks map inputs to outputs through layers of parameterized transformations. Feedforward networks, universal approximation results, and ImageNet-era convolutional networks are three stable anchors for understanding the basic idea.

## Core Explanation

This entry avoids broad unsourced claims about biological inspiration or current model performance. Its exported claims are limited to textbook feedforward-network structure, Cybenko's universal-approximation result, and the documented ImageNet impact of AlexNet.

## Further Reading

- [Deep Learning - Chapter 6, Deep Feedforward Networks](https://www.deeplearningbook.org/contents/mlp.html)
- [Approximation by Superpositions of a Sigmoidal Function](https://doi.org/10.1007/BF02551274)
- [ImageNet Classification with Deep Convolutional Neural Networks](https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks)

## Related Articles

- [Backpropagation: The Engine of Neural Network Learning](../backpropagation.md)
- [Mechanistic Interpretability: Reverse-Engineering Neural Network Circuits and Features](../mechanistic-interpretability.md)
- [Activation Functions in Neural Networks](../activation-functions.md)