2024 Nobel Prize in Physics awarded for neural network research

The Nobel Committee has acknowledged the contributions of two scientists to the advancement of artificial intelligence, awarding them the Physics Prize for the year 2024. John Hopfield and Geoffrey Hinton developed methods that form the foundation of modern machine learning.

The laureates utilized physical tools to create powerful data processing algorithms. Hopfield invented associative memory, capable of storing and retrieving images and other types of information patterns. Hinton developed a method that autonomously identifies properties within data, allowing tasks such as recognizing specific elements in images.

Artificial neural networks, inspired by the architecture of the brain, consist of nodes that represent neurons. These nodes interact through connections similar to synapsesA synapse is a link between neurons (or nodes in artificial neural networks) through which signals are transmitted., which can be strengthened or weakened. The network is trained, for instance, by reinforcing connections between nodes with simultaneously high values.

Hopfield's network employs a method for storing and retrieving patterns. It is based on physical principles that describe material properties through atomic spinAtomic spin is a quantum mechanical term describing the intrinsic angular momentum of a fundamental particle, such as an electron or atom.. The network is trained by finding values for the connections between nodes such that the stored images possess low energy. When a distorted or incomplete image is fed to the network, it gradually updates the node values, reducing system energy to find the most similar stored image.

Geoffrey Hinton utilized Hopfield's network as a basis for a new network employing a different method: the Boltzmann machine. It can learn to recognize characteristic elements within a given type of data. Hinton utilized tools from statistical physics—the study of systems comprised of many similar components. The machine is trained by presenting examples that are likely to occur during its operation. The Boltzmann machine can be used for classifying images or generating new examples of the type of patterns it was trained on. Hinton advanced this work, helping to initiate the current explosive development of machine learning.

“The work of the laureates has already brought tremendous benefits. In physics, we are using artificial neural networks in a vast number of fields, for example, in developing new materials with specific properties,” says Ellen Muns, chair of the Nobel Committee for Physics.

The Nobel Prize in Physics has a rich history. It was established by Alfred Nobel, the Swedish chemist and inventor of dynamite, in 1895. Among its past laureates are notable scientists who made revolutionary discoveries in quantum mechanics, nuclear physics, astrophysics, and other fields.