ARTIFICIAL NEURAL NETWORKS

• About the Bulletin
• Artificial Neural Networks
• Petroleum Updates

About the Bulletin
This bulletin is written by students of the Centre for Oil and Gas Technology at IPS.
The content and views expressed in this document reflect only the opinion of the authors and not of the School.

ARTIFICIAL NEURAL NETWORKS
By definition,   a  neural network is a  collection of hardware and software components that is modeled after the way neurons work in the human brain. Essentially, it enables computers to think and learn in the same way that humans do. Assume that you walk to work. Your decision to report to work is mostly influenced by the weather. Given the binary nature of the inputs, let us frame the conditions as yes or no questions.  Is the weather cooperating?  1 indicates yes, 0 indicates no.

This adjustment of our knowledge and comprehension of the universe is based on pattern recognition.  And, like people, computers too learn via pattern recognition. This learning is at the heart of how neural networks operate. Essentially,  it is a trained system that seeks out and adapts to patterns inside data. It is modeled precisely after the way our own brain functions.

Artificial Neural  Networks are made up of many Perceptrons (a  simplified version of a Neural Network), that are combined together to form a multiple-layer neural network.

Neural network (Source: Medium)

Real-life applications of Neural Networks

1. Speech recognition: Consider the Amazon Echo Dot — a true work of magic.
2. Handwriting recognition: Neural networks can be trained to recognize patterns in a person’s handwriting.

 Petroleum Updates