Course Content

# Introduction to Neural Networks

Introduction to Neural Networks

## Basics

Imagine that you want to learn how to translate text from English into Spanish. You learn languages by memorizing words and phrases, their meanings, and the context in which they are used. Based on this experience, you will be able to translate new texts that you have never seen before.

Another case is the classification of cats and dogs. Just as a person learns to distinguish them from examples seen in life, so a neural network can learn to distinguish them from such examples.

The neural network does something similar. It learns from examples - it can be texts, images, sounds, any data that we want it to process. A neural network, just like a person learns a language, tries to identify patterns in this data.

It then uses these patterns to perform tasks such as classification (determining which category an object belongs to), regression (predicting a numerical value such as the price of a house), or generation (creating new content based on the learned patterns). This process of training a neural network using examples is called supervised learning and this is the most common way to train it.

Note

When a neural network is trained, labeled examples are fed as input. When we want to get a prediction from it, inputs are not labeled.

Here's a demonstration of a Neural Network designed to identify drawings of cats and dogs. It addresses a classification problem by taking an input of an unknown class and producing an output with the identified class. Experience it firsthand to understand its functionality.

LMB (Left Mouse Button) - to draw.

Shift + LMB - to erase.

Join us in this learning journey, and we will guide you through the process of creating similar neural networks, step by step.

## Neural Network Structure

But what is the difference between learning languages and teaching a neural network? Well, the important thing to understand here is that a neural network is a structure made up of layers of "neurons" that are analogous to the biological neurons in the brain. Each neuron processes information, receives input, and passes the result on. The neurons within the layer work in parallel, processing information and passing the results to the next layer until we get the final answer to our question.

Neurons in a neural network have connections, each of which has a so-called “weight”. This weight shows how important the connection between the two neurons is.

Note:

The thicker the connection, the more important it is.

The process of training a neural network is to adjust the "weights" of each neuron in such a way that the results they give are the most accurate. It's a bit like how we learn to play a musical instrument, gradually improving our skills and accuracy.

However, it is important to understand that neural networks are only a tool, they do not have their own consciousness or understanding of the world, like a person. They simply process the data and find the patterns that we asked them to find. And a neural network trained to predict the price of a house would not be able to predict the price of a guitar in a music store.

1. What does Supervised Learning mean?
2. What is a neural network in general?