1. Deep Learning - Introduction to Neural Networks
Deep learning (also known as deep structured learning [Deep Structured Learning], hierarchical learning [Hierarchical Learning] or deep machine learning [Deep Machine Learning]) is a collection of algorithms and a branch of machine learning.
Deep learning methods have shown impressive accuracy in areas such as conversation recognition, image recognition, and object detection in recent years.
However, the term "deep learning" is very old. It was proposed by Dechter in the field of machine learning in 1986, and then introduced into artificial neural networks by Aizenberg et al. in 2000. And now, because Alex Krizhevsky won the ImageNet competition in 2012 using a convolutional network structure, everyone's attention.
The father of convolutional network: Yann LeCun
Deep Learning Demo
Link: http://playground.tensorflow.org
2. Each layer of deep learning is responsible for the content
Each layer of the neural network is responsible for the content:
Layer 1: Responsible for identifying colors and simple textures
Layer 2: Some neurons can recognize finer textures, cloth patterns, engraved patterns, leaf patterns, etc.
Layer 3: Some neurons are responsible for feeling yellow candlelight, high light, fireflies, egg yellow, etc. in the dark night.
Layer 4: Some neurons recognize the existence of cute dog faces, pet shapes, cylindrical objects, seven-star ladybugs, etc.
Layer 5: Some neurons are responsible for recognizing flowers, black-eyed animals, birds, keyboards, archetypal roofs, etc.