Just hearing the words “machine learning” may conjure visions of science fiction movies like The Terminator in which artificial intelligence is used to create robots with the ability to think and act like humans. But machine learning is more than just a novel concept; it's being used in real-world applications, ranging from data mining cyber security to healthcare and retail.
Machine Learning Explained
In case this is your first time hearing about machine learning, let me give you a brief explanation of the term. Machine learning is essentially a form of artificial intelligence in which computers have the ability to learn on their own. Computers have long been programmed with various algorithms for the purpose of identifying patterns and collecting data. What makes machine learning computers different, however, is their ability to apply this data. So instead of just harvesting/collecting data, they actually apply it to their existing algorithms, changing and evolving into an even smarter computer.
How Machine Learning is Being Used Today
Machine learning technology can be found in a vast array of industries and applications. Banks and financial institutions use it to detect fraud; search engines like Google and Bing use it to enhance the quality of their search results; retailers use it to provide prospective customers with credit scores and customized offers; Informations Technology (IT) teams use it to detect network intrusion; email service providers use it to catch and filter spam messages; and healthcare organizations use it to predict wait times in hospital emergency rooms.
Machine Learning and Data Mining
Both machine learning and data mining share some similarities in terms of methods, but there are also nuances between the two. Machine learning is aimed at predicting what comes next based on the training data, whereas data mining is aimed at the discovery of previously unknown data and/or patterns or properties within that data.
Machine Learning and Statistics
Michael L. Jordan, professor at the University of California, Berkeley and researcher in artificial intelligence, explains that machine learning shares similar methods, principles and theoretic tools as statistics. In a Quora Q&A, several users have chimed in giving their input on the differences between machine learning and statistics, many of whom acknowledge that while they are similar, machine learning focuses on the goal of solving a complex task by allowing the machine or computer to learn.
The Three Tasks of Machine Learning
Tasks involving machine learning can typically be broken into three basic categories:
Supervised learning – as you may have guessed, in supervised learning the computer is given example inputs and outputs by a human teacher. The teacher observes while the computer attempts to learn a correlation between the inputs and outputs.
Unsupervised learning – the computer must learn to find structure in its inputs without the use of labels. Unlike its supervised counterpart, there is no human teacher present to provide the computer with example inputs and outputs.
Reinforcement learning – the third and final type of machine learning, reinforcement, occurs when a computer program seeks to accomplish a specific goal in a dynamic environment without a teacher.
Thanks for reading and feel free to let us know your thoughts in the comments below regarding the benefits of Agile Software Development.