Neural Networks: The Machine Brain

Lotus Labs
2 min readSep 24, 2020

Our brain is made up of billions of nerve cells, called neurons. Arranged in a network, neurons communicate with each other in a sequence-like manner: Once one neuron receives information, it passes it to another neuron, which passes it to another and so on. Neurons can even signal several neighboring neurons at once. For example, one neuron might transmit information about the location of a baseball pitch that’s flying toward you while other neurons alert your muscles to get ready to swing the bat.This interconnected system allows us to constantly accumulate new information as the more familiar we become with a concept, the more efficient neural messaging between cells becomes. Once we master a skill, it takes minimal effort for the neurons to relay information to each other - the neurons essentially become wired together.

Neural Networks are not unique to humans as machines also use a similar system to learn particular tasks assigned to them. For machines Neural Networks are Deep Learning algorithms which help them learn more efficiently and improve their scores. In a machine Neural Network, neurons are arranged in three layers — Input, Hidden and Output. The Input layer directly accepts the real data values, and the Output Layer gives the final estimate of the output. These two layers are connected together by hidden layers which apply weights to the inputs and direct them through an activation function as the output. The hidden layer can have any number of layers. When there are more than two hidden layers, Neural Network becomes a Deep Neural Network. As we add more and more layers and neurons, the model detects more patterns in the data.

Diagram of Input, Hidden, and Output layers

Each neuron in a layer identifies patterns in the data and sends it to the neurons in the next layer until the signal eventually reaches the output layer. For example, let us envision a scenario where we give an image of a horse to the neural network model. Neurons in the input layer learn the pattern of the horse’s eyes, legs, and other body parts. This information is then passed on to the rest of the layers where it is stored as information. Therefore, when a new image is given to the model, the neurons can reference patterns which they have already learned and determine whether the image is a horse. Ultimately, each layer is dependent on the other, and together these three layers form a powerful problem-solving system capable of handling numerical data, images, videos and audios.

--

--

Lotus Labs

Transform your business into an AI-driven enterprise. We specialize in Machine learning for Retail, Insurance, and Healthcare industries. www.lotuslabs.ai