Deep Learning Course — Lesson 2: Types of Artificial Neural Networks

While we’ve established that an artificial neural network is composed of interconnected layers of nodes or “neurons”, it’s essential to note that there are different types of neural networks designed for various tasks. Here, we’ll discuss three foundational types:

1. Feedforward Neural Networks (FNN)

Feedforward Neural Networks are the simplest type of artificial neural network. In a feedforward network, the information moves in only one direction — from the input layer, through the hidden layers, to the output layer. There are no cycles or loops in the network; information is always fed forward, never back.

Use cases of FNN include sales forecasting, customer research, risk management, etc.

2. Convolutional Neural Networks (CNN)

Convolutional Neural Networks are primarily used for image processing, clustering, and recognition tasks. CNNs are designed to automatically and adaptively learn spatial hierarchies of features from tasks where the input is grid-like, e.g., an image.

The name “convolutional” indicates that the network employs a mathematical operation called convolution. Convolution is a specialized kind of linear operation, and it’s essential to image processing tasks.

3. Recurrent Neural Networks (RNN)

Recurrent Neural Networks are designed to recognize patterns in sequences of data, such as text, genomes, handwriting, or the spoken word. Unlike feedforward neural networks, RNNs can use their internal state (memory) to process sequences of inputs — this makes them exceptionally well-suited for tasks that involve sequential data.

Let’s test your knowledge on the different types of neural networks discussed. Here are the questions:

1. Describe the main difference between a feedforward neural network and a recurrent neural network.

2. Which type of neural network would you use for an image classification task? Explain why.

3. Suppose you were developing a model to predict the next word in a sentence. Which type of neural network would you use and why?

4. Explain why convolutional neural networks are particularly suited to image processing tasks.

5. What is the main characteristic of a recurrent neural network that makes it suitable for handling sequence data like time series or text?

6. Provide a brief description of what “feedforward” means in the context of neural networks.

7. Why might we choose to use deep learning models, such as the neural networks discussed in this chapter, over traditional machine learning models for complex tasks?

As before, take your time to consider the questions and answer in your own words. Feel free to comment.