Training Neural Networks in C++

Training Neural Networks in C++

2h 7mAdvanced2023-07-28

Authors

Eduardo Corpeño

Eduardo Corpeño

Electrical Engineer, Computer Programmer, and Teacher for 15+ years

Course details

Learn about the purpose, structure, and training process of neural networks to improve your machine learning skills. In this project-based course, instructor Eduardo Corpeño teaches you how to create an intelligent system with a neural network from scratch in C++, as well as how to choose the right neural network architecture and training method for each problem. Eduardo starts by explaining the difference between a neural network and other programming tools. He goes over why this course uses C++ and how to add different types of neural networks to your toolbox. The inspiration for artificial neural networks is the brain, and Eduardo relates parts of a biological neuron to C++ elements, then shows how to use activation functions and perceptrons in building neuron models. Eduardo covers the steps you will need to build and train your network. He explains segment display recognition, then guides you through designing and training your own SDR neural network.

Skills covered

C++Neural Networks and Deep LearningAdvancedArtificial Intelligence (AI)Programming LanguagesOpen SourceSoftware Development

Concepts

Introduction

  • Create a neural network from scratch in C++
  • What you should know

Choosing a Neural Network

  • What is a neural network
  • Why C++
  • The many applications of machine learning
  • Types of classifiers
  • Types of neural networks
  • Multilayer perceptrons

The Building Blocks of Neural Networks

  • Neurons and the brain
  • A simple model of a neuron
  • Activation functions
  • Perceptrons - A better model of a neuron
  • Challenge - Finish the perceptron
  • Solution - Finish the perceptron
  • Logic gates
  • Challenge - Logic gates with perceptrons
  • Solution - Logic gates with perceptrons

Building Your Network

  • Linear separability
  • Writing the multilayer perceptron class
  • Challenge - Finish the multilayer perceptron class
  • Solution - Finish the multilayer perceptron class

Training Your Network

  • The need for training
  • The training process
  • Error function
  • Gradient descent
  • The delta rule
  • The backpropagation algorithm
  • Challenge - Write your own backpropagation function
  • Solution - Write your own backpropagation function

Making a Segment Display Classifier

  • Segment display recognition
  • Challenge - Design your own SDR neural network
  • Solution - Design your own SDR neural network
  • Challenge - Train your own SDR neural network
  • Solution - Train your own SDR neural network
  • 7 to 1 network GUI demo
  • 7 to 10 network GUI demo
  • 7 to 7 network GUI demo

Conclusion

  • Next steps
80,000 Toman