Intro to Machine Learning (ML Zero to Hero, part 1)

Machine Learning represents a new paradigm in programming, where instead of programming explicit rules in a language such as Java or C++, you build a system which is trained on data to infer the rules itself. But what does ML actually look like? In part one of Machine Learning Zero to Hero, AI Advocate Laurence Moroney walks through a basic “Hello World” example of building an ML model, introducing ideas which we’ll apply in later episodes to a more interesting problem: computer vision.

Neural Network 3D Simulation

This neural network 3D simulation demonstrates how different models read visual images of hand-written numbers to translate and identify the images as their respective characters.

The iterative process of computational data comparison and pattern recognition is animated via 4 models:

  • Perceptron
  • Multilayer Perceptron
  • Convolutional Neural Network
  • Spiking Neural Network

The animated simulations offer an opportunity to observe the relatively abstract operations of algorithmic data processing.

What is a Neural Network?

A neural network is a computer system modeled on the human brain and nervous system and is trained to recognize patterns.

The patterns they recognize are numerical, but all real-world data, whether images, sound, text or time series, can be translated.

There are many neural network variants, but in this video introduction, a basic model is demonstrated.

When talking about neurons in a computer, it’s just something that holds a number.

A series of connected neurons comprise a layer.

And a neural network is a series of neuron layers.

Neural networks cluster and classify data.

They are used to group unlabeled data according to similarities among the example inputs, and they classify data when they have a labeled set of data for comparison.

In this video, the idea it to learn patterns from an image to read a hand-drawn number.

Much of machine learning comes down to having a good grasp of linear algebra.

In brief, this type of neural network is based upon different types of mathematical averages and biases that build upon recognizing parts of images to recognizing a complete image.