Comprehensive coverage

Nobel Prize in Physics 2024 for researchers on the development of artificial neural networks (extension)


John Hadfield and Geoffrey Hinton won the prize for their pioneering research in machine learning using neural networks inspired by the structure of the brain

2024 Nobel Prize in Physics winners John Joseph Hadfield of Princeton University and Geoffrey Hinton of the University of Toronto. Ill. Niklas Elmehed © Nobel Prize Outreach
2024 Nobel Prize in Physics winners John Joseph Hadfield of Princeton University and Geoffrey Hinton of the University of Toronto. Ill. Niklas Elmehed © Nobel Prize Outreach

The Nobel Prize in Physics for 2024 was awarded to two scientists equally: one scientist from the USA and another from Canada who studied neural networks.

The two laureates of the 2024 Nobel Prize in Physics are: John Joseph Hadfield of Princeton University and Geoffrey Hinton of the University of Toronto. The researchers received the prestigious award thanks to a long series of groundbreaking studies that enabled development in the field of machine learning and neural networks, which are used as calculation tools that simulate the activity of the nervous system.

The two researchers were able to train artificial neural networks using physics.

 

The two laureates of the Nobel Prize in Physics for this year took tools from the field of physics and with their help developed methods that are now the foundations of today's powerful machine learning.

John Hadfield from Princeton University developed an associative memory that can store and reconstruct images and other types of patterns within a database.

Jeffrey Hinton from the University of Toronto invented a method that can independently find features within a database, thereby performing tasks such as identifying specific elements in images.

 When we talk about artificial intelligence, we usually mean machine learning using artificial neural networks. This technology was originally inspired by the structure of the brain. In an artificial neural network, the neurons in the brain are represented by nodes with different values. These nodes influence each other through connections connected by synapses (couplers) that can be strengthened or weakened. The network is "trained", for example by developing stronger connections between nodes with higher values, at the same time. This year's Nobel laureates conducted important research with the help of artificial neural networks from the early 1980s onwards.

John Hadfield invented a grid that uses a method for saving and rebuilding patterns. We can imagine the nodes as pixels. The Hopfield network, a reliable feedback network, takes advantage of physics in which the properties of materials are described by the atomic spin - a property that turns each atom into a tiny magnet. The network as a whole is described in a manner analogous to the energy stored in the spin system known from the field of physics, and is trained by finding values ​​for the connections between the nodes so that stored images have low energy. When the Hopfield network is fed a distorted or incomplete image, it systematically goes through the various nodes and updates their values ​​so that the total energy of the network is reduced. Thus, the network works incrementally to find the saved images that are most likely to be incomplete after being fed.

Jeffrey Hinton of the University of Toronto used the Hopfield network as the basis for a new network that uses a different method: the Boltzmann machine (Boltzmann machine). This method can "learn" to identify characteristic elements in a given type of database. Hinton used tools taken from the field of statistical physics, the science in which systems are made up of many similar components. The machine is trained by feeding it examples that are most likely to be executed when the machine is turned on. A Boltzmann machine can be used to classify images or create new examples of the type of pattern with which it has been "trained". Hinton built on this work, and with it he pioneered the entire field of machine learning.

"The research of this year's Nobel laureates already has great benefits. In physics, we use artificial neural networks in a variety of fields, such as the development of new materials with required properties," says Ellen Moon, chair of the Nobel Committee in Physics.

· John Hadfield - born in 1993 in Chicago, Illinois, USA. He received his PhD in 1958 from Cornell University, a research university in Ithaca, New York, USA. Currently serves as a professor at Princeton University, New Jersey, USA.

· Geoffrey Hinton – born in 1947 in London, United Kingdom. Received his PhD in 1978 from the University of Edinburgh. Currently serves as a professor at the University of Toronto, Canada.

More of the topic in Hayadan:

Leave a Reply

Email will not be published. Required fields are marked *

This site uses Akismat to prevent spam messages. Click here to learn how your response data is processed.