John Wettlaufer and Hans Hansson on the 2024 Nobel Prize

John J. Hopfield delivering his Nobel Prize lecture in physics on 8 December 2024 at the Aula Magna, Stockholm University. | © Nobel Prize Outreach. Photo: Anna Svanberg

In his 1972 article “More is Different” Nobel Laureate Philip Anderson assails the reductionist hypothesis[1], and the immediate corollary; “that if everything obeys the same fundamental laws, then the only scientists who are studying anything really fundamental are those who are working on those laws.” He went on to demonstrate that, despite the wide acceptance of the reductionist hypothesis, an even more grievous implication is the validity of the “constructionist” hypothesis, which purports that one can reconstruct the universe beginning with the fundamental laws, which of course we now know is not possible.

In contemporary parlance, a complex system is one in which some aspects of the global-collective behavior are the consequence of the interactions between many simple constituents, but the behavior cannot be predicted solely from the rules underlying these interactions. This so-called “emergent behavior” is not amenable to explanation based solely on how the constituent parts interact. In this spirit, John Hopfield, one of this year’s Nobel laureates, wrote in his 1982 paper[2];

“In physical systems made from a large number of simple elements, interactions among large numbers of elementary components yield collective phenomena such as the stable magnetic orientations and domains in a magnetic system or the vortex patterns in fluid flow. Do analogous collective phenomena in a system of simple interacting neurons have useful "computational" correlates? For example, are the stability of memories, the construction of categories of generalization, or time-sequential memory also emergent properties and collective in origin? This paper examines a new modeling of this old and fundamental question and shows that important computational properties spontaneously arise…The model could be readily implemented by integrated circuit hardware”

He then presented a model for a stable associative memory, based on a network of nodes, or “neurons”, that are connected by links, or synapses, of adjustable strength. Using reasoning based on an impressive panoply of concepts in physical theory, he demonstrated that memories can be retrieved starting from many different states. This is reminiscent of how our own memories can be triggered by different stimuli like a smell, a picture or a tune.

Since then, artificial neural networks have been applied to a myriad of areas, such as visual pattern recognition, symbolic computation, and the translation and interpretation of texts. Specifically, they have also become important tools for research in both physics and other natural sciences.

Applications aside, this year’s prize is for foundational work showing how complex patterns and functions can emerge spontaneously in a myriad of physical systems, ranging from mammalian brains to electronic circuits. Both the foundational topics and applications are studied at Nordita, particularly within the framework of the Wallenberg Initiative on Networks and Quantum Information—WINQ.

Thors Hans Hansson and John Wettlaufer

[1] P.W. Anderson, More Is Different: Broken symmetry and the nature of the hierarchical structure of science. Science 177, 393 (1972).

[2] J.J. Hopfield, Neural networks and physical systems with emergent collective computational abilities. Proc. Natl Acad. Sci. USA 79, 2554 (1982).