Can computers really learn? How AI learns
In 2016, an AI beat a top Go player 4 to 1!
AI does not follow a rule a person wrote for every step. It learns from examples through artificial neural networks modelled on the brain: when it guesses wrong, it adjusts itself until it gets better and better.
In 1956 AI became a new field at the Dartmouth workshop. In 1958 the Perceptron learned to tell left-marked cards from right-marked ones after 50 tries. In 2012 AlexNet used graphics chips to learn from a huge set of images. And in 2016 AlphaGo beat a top Go player 4 to 1.
Sources: Google DeepMind, Computer History Museum, Britannica, IEEE Spectrum, Cornell University
- #ບອກຕໍ່ຄວາມຮູ້
- #ຮູ້ບໍ່EP10
- #ເຕັກໂນໂລຢີ
- #AI
- #MachineLearning
- #shorts
Watch and share on
Watch next
The first computer bug was a real insect
The very first computer bug was an actual insect!
Python: the language of AI was born from a hobby
The language of AI was born from a Christmas hobby in 1989!
Venus: a day longer than a year
Did you know there is a planet where one day is longer than one year?
Mona Lisa: famous because it was stolen?
Did the world's most famous painting become famous because it was stolen?