Skip main navigation

£199.99 £139.99 for one year of Unlimited learning. Offer ends on 28 February 2023 at 23:59 (UTC). T&Cs apply

Find out more

Artificial neural network (ANN)

What is Artificial Neural Network?
Hello everyone. Now in this video, I will explain about the artificial neural network. Neural network is the building block of deep learning, which is a sub field of machine learning, where the algorithm inspired by the structure of the human brain.
You can see in the left image, a human real neuron, which has dendrites which receive information from other neurons. And the axon, which transfer these information to another neuron. And you can see the synaptic weight these in a yellow color, which store information in our neuron. Our goal, we want. to make the computer intelligent, we want to simulate / to mimic this intelligence(brain intelligence) in machine. So our goal is to mimic the neural network. You can see here we have different inputs in the left image, . This is the building block of the artificial neruon.
We have inputs: X1, X2, to Xn and these inputs also have synaptic weights (W), then it outputs Y This is how we simulate a real neuron.
What artificial neural dose? it sums the multiplication of the input with the weights, (with their corresponding weights), then it goes through some activation function. You don’t need to know the details. I just want to give you a high level introduction to artificial neural network.
We can see now in this figure, we have artificial neural network. We have our input layer and one hidden layer and one output layer, the hidden layer containes five neurons. So each circle is a neuron and we have one output neuron output neuron in the output layer. We may have a deeper network with more than one hidden layer, maybe three hidden layers, seven layers or more hundreds hidden layers. And also, maybe we can use the neural network for classification. As you can see here, we have three neurons in the output layer, each neuron for each class. We have here three classes. Now, we will take an example, a very simple example.
We have an image and the pixels for that image, we feed the image(features) to our neural network. You can see here we have inputs neurons, hidden neurons and output neurons. You can see in the output layer, we only have three neurons. Why? because we only have three classes. Our network want to classify the input image into one of three classes. Is this image a bicycle or a train or a car. So we have only three neurons in the output layer.
First, we fed our features of our image to our network on our network should be classification(network) to do classification for that image. The input for this network; in this example is a human designed features, like number of circle, What is the biggest shape? What is the number of pixels, color of the pixels? Then we feed these features to the input layer, then the network will process these data and after that, we will get our classification, which is a train. We also can feed our network instead of this features, Just simply feed our network the pixels of that image. So this network is learned by training.
After that, we use the network(trained network) to predict the class of a new image, which is train. Now, this video is really a nice video. This is an optional task, you can try it by yourself. Just the click to that link on the try to see by yourself a symbolic classification problem. For example, here we have to select our data, I selected the data I want to classify. Then I can simply control the number of hidden layers number of neurons, in input layer, number of output, number of the neuron. Maybe choose
the activation function. You can see different problem from the data. Here I reduced the number of the input and increased the number of the hidden layers. Then I will start the simulation. See, it’s not really easy to classify. Let’s try to change the activation function. This is the function we talk about in the first slides. I will change it to a ReLu function. Now you can see we can classify very fast. You can try it by yourself just play with all these options, increase number of neurons, different activation function, choose different data to be more familiar with the concept of artificial neural network. Thank you.

You may ask yourself why this network is called an artificial neural network?

  • It is called that because the building block for it (the perceptron) looks like/mimics the neurons in the brain.

In this video, you will be introduced to the ANNs at a high level and see how they work.

The video will take you through an interactive visualisation of neural networks which contains a tiny neural network library that meets the demands of this educational visualisation. You can simulate, in real-time in your browser, small neural networks and see the results.

Discussion :

  • Can you explain in simple words what artificial neural networks are?
This article is from the free online

An Introduction to Artificial Intelligence in the Tourism Industry

Created by
FutureLearn - Learning For Life

Our purpose is to transform access to education.

We offer a diverse selection of courses from leading universities and cultural institutions from around the world. These are delivered one step at a time, and are accessible on mobile, tablet and desktop, so you can fit learning around your life.

We believe learning should be an enjoyable, social experience, so our courses offer the opportunity to discuss what you’re learning with others as you go, helping you make fresh discoveries and form new ideas.
You can unlock new opportunities with unlimited access to hundreds of online short courses for a year by subscribing to our Unlimited package. Build your knowledge with top universities and organisations.

Learn more about how FutureLearn is transforming access to education