Machine learning is an algorithm that uses statistics to analyze large amounts of digital data such as numbers, words, and images. It revolves around the principle of find the pattern, apply the pattern. Machine learning is an almost omniscient presence in our daily lives as it powers recommendation systems, search engines, social media feeds, and voice assistants on digital services such as YouTube, Baidu, Instagram, and Siri. These platforms make educated guesses on what a user would want next or, in the case of voice assistants, which words best match the words a user speaks by collecting data such as the links being clicked, searches being frequently made, and the reaction used.
Machine learning began in 1986 with the invention of deep learning which uses a technique that enhances a machine’s ability to detect and amplify even the smallest patterns. The technique in question is called a deep neural network that operates through many layers of computational nodes that work together to sift through data and find a pattern. Neural networks in many ways resemble the human brain where the nodes are the neurons and the network is the brain itself.
Both machine and deep learning employ supervised, unsupervised, and reinforced methods. Supervised learning machines are told exactly what patterns to seek; when pressing play on a Netflix show, the algorithm is directly told to look for similar shows thereafter. Unsupervised learning is when the algorithm looks for whatever patterns it can find and while it has less applications, it has become popular in cybersecurity. Reinforcement learning achieves an objective through a trial and error process; many different methods of pattern-detection are tried and based on rewards and penalties, the algorithm decides whether its approach is beneficial or not to achieving its goal.
Karen Hao. (November 2018). What is machine learning? Retrieved from https://www.technologyreview.com/2018/11/17/103781/what-is-machine-learning-we-drew-you-a nother-flowchart/
Comments