# 9. Convolutional Neural Networks

## CNNs in less than 100 words
````{margin}
Applications of CNNs are not restricted to image inputs: they are used with audio data for voice recognition, as well as text data in the field of Natural Language Processing (a collection of techniques and approaches that are used to analyze and understand natural language data).
````
Convolutional Neural Networks are specific artificial networks reigning supreme in image processing tasks such as image classification, object detection and image generation.

CNNs contain special layers that take into account the neighbouring relationships between pixels, thus preserving the spatial structure in an image data. CNNs work by applying a series of filters to the input image to extract relevant features such as edges, shapes, and textures. They use these features to make predictions or take actions.

## This lecture

This lecture will give you the basics of CNNs, the key components of their architecture and how they work.

```{admonition} Special Format!
:class: note

This lecture has a specific format. It is composed of two parts. The first one is a "take home lecture." I selected some __videos__ for you to watch so that you can __learn at your own pace__. For some videos you will appreciate the wonderful animations that talented YouTubers create. 

The second part is a __guest lecture__ by Dr. Ben Nachman, an experimental particle physicist and machine learning expert, who will talk about specific applications of CNNs in the context of particle physics. No technical knowledge in the subatomic world is required to understand. The idea is to expand your knowledge with implementations departing from all the classical examples one can encounter in the majority of textbooks and online resources.

Happy learning!
```