In June of 1956, a small workshop of less than a dozen scientists and mathematicians met at Dartmouth College in Hanover, New Hampshire. The Dartmouth Summer Research Project on Artificial Intelligence not only created a new field of research, it founded what became a worldwide industry, today funded by governments and corporations to the tune of hundreds of billions of dollars.
The terms “artificial intelligence,” “machine learning,” and “deep learning” have become technology and fiction buzzwords, frequently appearing in science news, advertising, movies, and science fiction stories. What exactly is artificial intelligence, and how do the fields of machine learning and deep learning relate to it?
Artificial Intelligence or AI is a broad-based, interdisciplinary field of study pursuing machines that are capable of performing tasks which require human intelligence to complete. Machine learning and deep learning are just two of the ways to equip machines with the knowledge necessary to act as if intelligent, in the mold of human thought and behavior.
More specifically, machine learning is the term for teaching a computer to perform a task, instead of programming a series of precisely ordered steps for the machine to follow. There are two main methods. Supervised learning labels types of data for the computer to use as examples. Unsupervised learning has the computer sort data into similar types, then spot detailed differences. This is how machines are taught tasks like facial recognition or predicting stock market trends.
Deep learning is the term for a specific type of machine learning that allows computers to understand complex problems and provide insight, solutions, or controls for those problems. This process involves the use of neural networks, which are groups of separate computer programs that perform specific computations, then output the results to another component of the network.
The term “deep” is a reference to the way neural networks can be layered to receive and transmit results to each other; exponentially multiplying the speed and complexity of machine learning. Deep learning produces results like following speech commands or recognizing the necessary cues to drive a car. This concept is what has allowed amazingly human-like feats of deduction and strategy accomplished by some computers, such as defeating chess grandmasters or predicting weather patterns.