a layout of neural networks
The ability of our brain to learn is reflected in our intellect. Artificial intelligence is used in computer systems that behave like humans. This means that these systems are managed by computer programs that can learn. Computers, like humans, may learn to utilize data and then make judgments or evaluations based on what they’ve learned. It is part of the broader area of artificial intelligence and is known as machine learning.
People wrote step-by-step instructions for the programs that control a computer’s hardware for computers to solve issues. Those programmers had to think about every move a computer would or could take. Then they outlined how they intended the machine to respond to any decisions that came up along the route.
Arthur Samuel chose to program computers differently in the 1940s while working as an engineer at Illinois. This computer scientist’s goal would be to educate computers on how to learn on their own. Checkers are his teaching instrument. Rather than programming every potential move, he let elite checkers players advise the machine. Consider these to be broad guidelines.
He also taught the computer how to play checkers with itself. The computer kept note of which of its plays and plans were the most effective during each game. Then it applied those movements and techniques to improve its performance the next time. The computer converted bits of data into information along the way. That data would create knowledge, allowing the machine to make better decisions. Within a few years, Samuel had created his first computer program to play that game. He was working in an IBM laboratory in Poughkeepsie, New York, at the time.
Increasingly Complicated Networks
Programmers quickly abandoned checkers. They trained computers to perform progressively complicated problems using the same method. In 2007, Fei-Fei Li and her colleagues at Stanford University in California sought to teach machines to detect things in photographs. We may think of sight as something we can only see with our eyes. In truth, our brains are responsible for recognizing and comprehending what an image depicts.
Li’s team fed enormous amounts of data into computer models. The computer required many images. And the researchers had to ensure that each image of a cat that the computer was trained on depicted a cat. Li explained her team’s approach in a 2015 TED presentation. They required the assistance of other scientists. It took roughly three years for nearly 49,000 volunteers from 127 countries to filter through over 1 billion photos.
Li’s team eventually came up with a collection of almost 62,000 pictures, all of which were of cats. It was ready for machine learning.
Some felines sat. Others took a stand. Or huddled. Or curled up in a ball. The images portrayed a wide range of animals, from lions to housecats. As machine algorithms combed through the data in these photos, they learned how to recognize a cat in every new shot that was given to them.