Neural Network Course Schedule

Introduction to NeuroSolutions and Neural Networks

We use NeuroSolutions for all of the demonstrations in our course not just because we created the software, but because it’s an excellent platform for experimenting with different techniques and topologies. This section of the course covers the fundamentals of neural networks and how to interact with NeuroSolutions to try the experiments. We recommend this section for everyone not familiar with NeuroSolutions.

Introduction to neural networks

• Terminology
• Fundamental principles of neural networks
• Overview of neural network architectures and training
• When to use and why should you use neural networks

Fundamentals of NeuroSolutions

• Overview of breadboards, palettes, families, etc.
• The Neural Wizard
• Placing and interconnecting components by hand
• Special focus on how to use probes, properties, and file components

Using NeuroSolutions

• Creating, training, and testing neural networks
• Using probes to understand the training process and the results
• What to look for and how to use probes in a network
• Using and setting the network parameters

Using NeuroSolutions for Excel

• Preprocessing and analyzing your input data
• Tagging your data
• Creating a neural network
• Training a neural network
• Testing a neural network
• Analyzing your results
• Optimizing neural network parameters / inputs

Overview of other features of NeuroSolutions

This section gives a broad overview of many of the common techniques that fall under the conceptual umbrella of neural networks. After completing this section, you should have a very good idea of what neural networks can and can’t do, what types of topologies work best for different types of problems, and how to get the most out of your data. It also highlights one of the most popular type of neural network problems: financial forecasting.

Fundamentals of Adaptive Systems and Neural Networks

• Adaptive Systems and Linear Regression
• Analyzing linear adaptive systems
• Understanding gradient descent training

Supervised Learning

• Overview of MLPs (nonlinear extenstions to linear adaptive systems)
• Tips and tricks of the trade: MLP parameters and how to set them
• Applications of MLPs
• Genetic optimization of parameters
• Project 1: Using MLPs for classification

Unsupervised Learning

• Intro to unsupervised learning
• Hebbian learning and principal component analysis
• Competitive learning and clustering (including SOMs)

Radial basis functions (RBFs)

• Introduction to unsupervised learning
• What are RBFs and why/when should you use them?
• How to use RBFs and how to set their parameters
• Hybrid unsupervised/supervised networks
• Project 2: Using a hybrid RBF/MLP for classification

Temporal processing and dynamical systems

• Adaptive signal processing fundamentals
• Temporal neural networks

• Optimizing inputs, learning rates, network size, etc. 

Financial Forecasting using Neural Networks

• Introduction to prediction and the stock market 
• Optimal trading signals
• Neural network prediction
• Building a trading system
• Analyzing and optimizing the trading system 

Overview of using the advanced features and capabilities of NeuroSolutions

• Using Macros to automate tasks 
  • Introduction to macros and the MacroWizard utility
  • Recording a sequence of events
  • Using the MacroWizard editor and debugger 
  • Assigning macros to dialog components and toolbar buttons
• Customizing components using DLLs
  • Creating a new processing element activation function
  • Updating the backpropagation plane
  • Creating a new error criteria
  • Creating a new gradient search component
  • Specialized I/O
    • Creating a new file translator
    • Reading/Writing data to/from an external source