Machine learning, a branch of computer science, developed from the study of pattern recognition and computational learning theory in artificial intelligence. In short, according to Arthur Samuel (1959), machine learning is a “Field of study that gives computers the ability to learn without being explicitly programmed.” It explores the pattern in algorithms so that can make predictions on data. By building a model from example inputs, such algorithms build a model in order to make data-driven predictions or decisions, instead of being “told” to do something. As a method used to design complicated models and algorithms that make predictions, machine learning is known as predictive analytics in commercial use. Therefore, such analysis helps people to better understand the data and come to reasonable and reliable decisions.

Once an academic discipline of artificial intelligence, machine learning that lets machines learn from data, interested some researchers. They used symbolic methods, “neural network”, which are later found as reinventions of the general linear models of statistics, and probabilistic reasoning in automated medical diagnosis. However, the increasing emphasis on the logical, knowledge-based approach in machine learning, differentiates it from artificial intelligence. It was reorganized as a separate field and began to flourish in the 1990s. The field stopped achieving artificial intelligence to solvable issues of a practical nature. Rather than inheriting he symbolic approaches from AI, it focused on methods and models borrowed from statistics and probability theory. The widespread use of digital information and the Internet benefited the development of machine learning a lot.

As a learner would generalize from its experience, in the field of machine learning, a learning machine should possess an ability to perform accurately on new problems or examples after having experienced a learning data set. The computational analysis of machine learning algorithms and their performance evolves from computational learning theory. Since training sets are finite and the future is uncertain, learning theory cannot guarantee the correctness of all calculations. Computational errors are quite common.

Machine Learning has been applied to various fields, including adaptive websites, brain-machine interfaces, computational anatomy, game playing, marketing, search engines, software engineering, cheminformatics, natural language processing and etc. As a branch of computer science, machine learning even influences the business world. In 2010, The Wall Street Journal wrote about money management firm Rebellion Research’s use of machine learning to predict economic movements, describing the company’s prediction of the financial crisis and economic recovery.



How can machine make predictions from itself?
What’s the limitation of machine learning at current technological level?
How will machine learning impact the industry?