Python Artificial Intelligence Projects for Beginners

deep learning
, and apply what is known as a
convolutional neural network to identify handwritten mathematical symbols. Finally we
will revisit the bird species identifier covered previously and use deep learning to produce
a much more accurate identifier.
The topics that we will be covering in this chapter are as follows:
Understanding neural networks
Identifying the genre of a song using neural networks
Recalling our work on the spam detector to use neural networks
Understanding neural networks
Neural networks, which were originally called artificial neural networks, are inspired by
actual neurons found in animal's brains and other parts of the nervous system. Neurons are
connected to each other and they receive and send impulses throughout the animal's body,
or in the case of computing, the network.

---

Neural Networks
Chapter 4
[ 78 ]
The following diagram shows the components of a single neuron:
%QORQPGPVUQHCUKPINGPGWTQP
The following graph shows how a neuron fires:
*QYCPGWTQPaTGU

---

Neural Networks
Chapter 4
[ 79 ]
It is all or nothing, meaning, when the neuron gets enough input from its neighbors, it
quickly fires and sends a signal down its axon to each forward-connected neuron.
Here, we can see actual neurons in a brain:
#EVWCNPGWTQPUKPCDTCKP
A human brain has about 100 billion neurons all together, and has about 100 trillion
connections. It is worth noting that the neural networks we create in software have at least
1 million times less complexity.

---

Neural Networks
Chapter 4