The Nobel Prize in Physics was awarded to John Hopfield and Geoffrey Hinton for their research in the field of AI: their scientific work will help develop machine learning using neural networks, which will make computers smarter and they will be able to better recognize pictures and speech.
As the Nobel Committee explained, the neural networks that are used in AI were created by analogy with the human brain; In such networks, neurons are represented by nodes, and their connections change in the process of learning, making it possible to solve complex problems, for example, to recognize patterns. Hopfield came up with a model where nodes behave like pixels, and the network resembles the behavior of atoms, turning them into small magnets. This network can repair corrupted images by constantly changing the values of the nodes until they get a picture that is as similar as possible to the original. Hinton, based on Hopfield's ideas, created the Boltzmann Machine, a network that learns to recognize important features in data. He used statistical methods to teach it to understand the examples, so the Machine can sort the images or even create new ones based on the ones it has already seen and studied.
