Training of physical neural networks

Ali Momeni, Babak Rahmani, Benjamin Scellier, Logan G. Wright, Peter L. McMahon, Clara C. Wanjura, Yuhang Li, Anas Skalli, Natalia G. Berloff, Tatsuhiro Onodera, Ilker Oguz, Francesco Morichetti, Philipp Hougne, Manuel Gallo, Abu Sebastian, Azalia Mirhoseini, Cheng Zhang, Danijela Marković, Daniel Brunner, Christophe Moser, Sylvain Gigan, Florian Marquardt, Aydogan Ozcan, Julie Grollier, Andrea J. Liu, Demetri Psaltis, Andrea Alù and Romain Fleury ()
Additional contact information
Ali Momeni: Ècole Polytechnique Fédérale de Lausanne (EPFL)
Babak Rahmani: Microsoft Research
Benjamin Scellier: Rain AI
Logan G. Wright: Yale University
Peter L. McMahon: Cornell University
Clara C. Wanjura: Max Planck Institute for the Science of Light
Yuhang Li: University of California, Los Angeles
Anas Skalli: CNRS UMR 6174, Institut FEMTO-ST, 25000
Natalia G. Berloff: University of Cambridge
Tatsuhiro Onodera: Cornell University
Ilker Oguz: Ècole Polytechnique Fédérale de Lausanne (EPFL)
Francesco Morichetti: Politecnico di Milano
Philipp Hougne: Université de Rennes, CNRS, IETR – UMR 6164
Manuel Gallo: IBM Research Europe – Zurich
Abu Sebastian: IBM Research Europe – Zurich
Azalia Mirhoseini: Stanford University
Cheng Zhang: Microsoft Research
Danijela Marković: CNRS, Thales, Université Paris-Saclay
Daniel Brunner: CNRS UMR 6174, Institut FEMTO-ST, 25000
Christophe Moser: Ècole Polytechnique Fédérale de Lausanne (EPFL)
Sylvain Gigan: Laboratoire Kastler Brossel, Sorbonne Université, École Normale Supérieure, Collège de France, CNRS
Florian Marquardt: Max Planck Institute for the Science of Light
Aydogan Ozcan: University of California, Los Angeles
Julie Grollier: Laboratoire Albert Fert, CNRS, Thales, Université Paris-Saclay
Andrea J. Liu: University of Pennsylvania
Demetri Psaltis: Ècole Polytechnique Fédérale de Lausanne (EPFL)
Andrea Alù: City University of New York
Romain Fleury: Ècole Polytechnique Fédérale de Lausanne (EPFL)

Nature, 2025, vol. 645, issue 8079, 53-61

Abstract: Abstract Physical neural networks (PNNs) are a class of neural-like networks that make use of analogue physical systems to perform computations. Although at present confined to small-scale laboratory demonstrations, PNNs could one day transform how artificial intelligence (AI) calculations are performed. Could we train AI models many orders of magnitude larger than present ones? Could we perform model inference locally and privately on edge devices? Research over the past few years has shown that the answer to these questions is probably “yes, with enough research”. Because PNNs can make use of analogue physical computations more directly, flexibly and opportunistically than traditional computing hardware, they could change what is possible and practical for AI systems. To do this, however, will require notable progress, rethinking both how AI models work and how they are trained—primarily by considering the problems through the constraints of the underlying hardware physics. To train PNNs, backpropagation-based and backpropagation-free approaches are now being explored. These methods have various trade-offs and, so far, no method has been shown to scale to large models with the same performance as the backpropagation algorithm widely used in deep learning today. However, this challenge has been rapidly changing and a diverse ecosystem of training techniques provides clues for how PNNs may one day be used to create both more efficient and larger-scale realizations of present-scale AI models.

Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.nature.com/articles/s41586-025-09384-2 Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:nat:nature:v:645:y:2025:i:8079:d:10.1038_s41586-025-09384-2

Ordering information: This journal article can be ordered from
https://www.nature.com/

DOI: 10.1038/s41586-025-09384-2

Access Statistics for this article

Nature is currently edited by Magdalena Skipper

More articles in Nature from Nature
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().