As the name Artificial Neural Networks (ANNs) suggests, the design of ANNs is inspired by the human brain. Although not as powerful as the brain (yet), artificial neural networks are the most powerful learning models in the field of machine learning.
In the past few years, deep artificial neural networks have proven to perform surprisingly well on complex tasks such as speech recognition (converting speech into text), machine translation, and image and video classification. Such models are also commonly called deep learning models.
Let’s begin our journey into deep learning with an introduction to artificial neural networks.
Artificial neural networks are said to be inspired by the structure of the human brain. Let’s first learn about the basic structure of the brain and the anatomy of a neuron and understand how information travels through neurons.
In simple words, the biological neuron works as follows: it receives signals through its dendrites, which are either amplified or inhibited, as they pass through the axons to the dendrites of other neurons.
Let’s now take a look at how an artificial neural network is similar to the human brain.
To summarise, the main bottleneck in using neural networks is the availability of abundant training data. Neural networks find applications across various domains such as images and videos (computer vision), text and speech. Note that the words ‘deep learning’ and ‘neural networks’ are often used interchangeably.
Also, artificial neural networks are a collection of many simple devices called artificial neurons. The network ‘learns’ to conduct certain tasks, such as recognising a cat, by training the neurons to ‘fire’ in a certain way when given a particular input, such as a cat. In other words, the network learns to inhibit or amplify the input signals to perform a certain task, such as recognising a cat, speaking a word or identifying a tree.
In the next segment, you will study the basics of a perceptron. The perceptron was one of the earliest proposed models for learning simple classification tasks, which later became the fundamental building block of artificial neural networks.
Report an error