90,150 views
Learn what neural networks are step-by-step, understanding how the learning process of a neural network works. Full course: https://didatica.tech/curso-redes-neu... This class is organized as follows: 01:39 Inspiration from the biological neurons of the human brain 03:40 What information does a neural network receive as input 04:50 What the neural network is trying to learn 07:05 Understanding what pixels in an image are 10:29 Introducing a simple neural network 12:55 The weights that multiply the inputs of a neural network 14:23 The activation function 16:03 The connections of the second neuron 17:43 How the neural network reports its answer in the output for a classification problem 20:30 The importance of network calibration 23:54 Why some weights are negative 28:18 How many weights and variables need to be calibrated in total 30:25 How a neural network learns (cost function and gradient descent) 49:13 How to test if the neural network is performing well (accuracy with test data) 50:01 Neural networks with more than one layer (deep learning) 59:19 What each layer is learning 01:05:00 Next steps to advance the study of neural networks and deep learning In this article you can check out more details on the subject: https://bit.ly/redeNeural Basically, a neural network is a structure that brings together small computing units (called neurons) in an organized way, allowing the combination of calculations performed by each neuron to result in solutions to complex problems. Each neuron in a dense (fully connected) neural network performs simple mathematical operations of addition and multiplication, and in these operations, many variables need to be calibrated. A neural network usually starts its training with random weights and biases, and quickly learns how to calibrate these variables until the output response is the desired one for classification or regression problems. To learn more about how the machine learning process works, read this article: https://bit.ly/oqueeML #neuralnetworks #deeplearning