Artificial intelligence research focuses on designing computer systems to mimic human intelligence by having a computer program make decisions or take actions based on the information provided.
What is Machine Learning?
Artificial intelligence research focuses on designing computer systems to mimic human intelligence by having a computer program make decisions or take actions based on the information provided. Artificial intelligence includes expert systems and machine learning; machine learning includes a subfield known as deep learning, which uses multiple layers, as opposed to “shallow learning.”
Research in artificial intelligence goes as far back as the 1950s, but was limited by computing hardware. In 1997, IBM’s Deep Blue demonstrated the possibilities of machine learning by defeating the world chess champion (Bernard Marr, “A Short History of Machine Learning—Every Manager Should Read,” Forbes , 2016, https://bit.ly/3f6UKWk ).
Although increased data storage capacity and computer processing power were helpful in advancing machine learning, two important developments in computers from the 1990s into the 2000s that accelerated the development of machine learning were the availability of large high-quality data and the development of parallel graphic processing units (GPU). Because machine learning requires large data sets in order to train the learning algorithms, the vast quantity of high-quality publicly available data allowed researchers to refine the machine learning algorithms. Parallel processing units were initially focused on meeting the demands of the graphic-intensive games, but the developments in parallel processing allowed significant enhancement of power for machine learning. Research and development in machine learning has rapidly accelerated in the past 15 years.
How Does Machine Learning Work?
Computer programs were initially created to give computers instructions to follow in solving a problem. This process, known as top-down programming, was used from the early days of computers until the 1990s when object-oriented programming (OOP) was created. OOP changed programming from isolated instructions to the computer to manipulate data, to treating the programs and the data that it manipulates into a defined object. This paradigm shift resulted in the rapid development of graphically based programs that were much easier to maintain because the programs were based on a set of self-contained objects that interacted with each other. OOP worked very well for typical programs such as word processing or spreadsheets.
The goal in machine learning is to write an algorithm that can be trained using test data to look for specific patterns.
The traditional methods of programming did not work very well for machine learning because they require a programmer to give the computer precise instructions for encountering both expected and unexpected possibilities. In machine learning, the expectation is that the algorithm will learn from the data provided, in a manner similar to how humans learn from data.[…]