Intramural call to promote research IIS Aragón 2024
October 8, 2024
Nobel Prize in Chemistry rewards AI revolution in protein research
October 10, 2024
The Royal Swedish Academy of Sciences has awarded researchers John Hopfield, from Princeton University (USA), and Geoffrey Hinton, from the University of Toronto (Canada), for their discoveries and inventions that enable machine learning with artificial neural networks.
This year's two Nobel Prize winners in Physics have used tools from physics to develop methods that underlie the powerful automatic learning current, one of the branches of the artificial intelligence (AI).
The Royal Swedish Academy of Sciences announced their names on Tuesday: John Hopfield, which created an associative memory capable of storing and reconstructing images and other types of patterns in data, and Geoffrey Hinton, who invented a method capable of autonomously finding properties in data and thus performing tasks such as identifying specific elements in images.
Both are rewarded “for foundational discoveries and inventions “enabling machine learning with artificial neural networks,” according to the Nobel Committee for Physics. This field is revolutionizing science, engineering and everyday life.
From natural to artificial neural networks
When we talk about AI we usually refer to the automatic learning through artificial neural networks. This technology was originally inspired by the natural neural structure and networks of the brain.
Natural neural networks have neurons that send signals through the process of synapses. When we learn things, the connections between some neurons become stronger, while others become weaker.
In an artificial neural network, those brain neurons are represented by nodes., which are encoded with different values. These are connected to each other and also influence each other through connections (such as synapses), which can strengthen or weaken each other.
The network is trained, for example, by developing stronger connections between nodes that have high values simultaneously. This year's winners have done important work with this type of artificial neural networks since the 1980s.
The Hopfield network
John Hopfield of Princeton University (USA) invented a network that uses a method to store and recreate patterns. We can imagine the nodes as pixels. The so-called Hopfield network uses physics that describes the characteristics of a material due to its spin atomic, a property that turns each atom into a tiny magnet.
The network as a whole is described equivalently to the energy in a spin system within physics, and is trained by finding values for the connections between nodes so that the saved images have low energy.
When the Hopfield network receives a incomplete image or distorted, it methodically goes through the nodes and updates their values so that the network's energy decreases. In this way, the network works step by step to find the saved image that most resembles to that imperfect one.
This year's #NobelPrize laureate in physics John Hopfield created an associative memory that can store and reconstruct images and other types of patterns in data.
- The Nobel Prize (@NobelPrize) October 8, 2024
The Hopfield network can store patterns and has a method for recreating them. When the network is given an… pic.twitter.com/QDDKymJCaF
Hinton uses the Boltzmann machine
Professor Geoffrey Hinton of the University of Toronto (Canada) used the Hopfield network as the basis for a new network that uses a different method: the Boltzmann machine. It can learn to recognize characteristic elements in a given type of data. Hinton used tools from statistical physics, the science of systems built from many similar components.
The machine is trained by feeding it examples that are very likely to appear when it is started. The Boltzmann machine can then be used to classify images or create new examples of the type of pattern it has been trained on. Hinton has built on this work, helping to initiate the explosive current development of the automatic learning.
2024 physics laureate Geoffrey Hinton used a network developed by his co-laureate John Hopfield as the foundation for a new network: the Boltzmann machine. This can learn to recognize characteristic elements in a given type of data.
- The Nobel Prize (@NobelPrize) October 8, 2024
The Boltzmann machine can be used to classify… pic.twitter.com/LMinR0vA0n
“The work of the two laureates has already been most beneficial. In physics we use artificial neural networks in a wide range of areas, such as the development of new materials with specific properties," he stressed. Ellen Moons, Chair of the Nobel Committee for Physics.
Hopfield, an American born in Chicago in 1933, received his PhD in 1958 from Cornell University (New York) and is now an emeritus professor at Princeton. Hinton, born in London (United Kingdom) in 1947, received his PhD in 1978 from the University of Edinburgh and, after moving to Canada, currently works at the University of Toronto.
Source: SINC Agency
Rights: Creative Commons.
Image: The 2024 Nobel Prize winners in Physics (Nobel Prizes)
