Probabilistic photonic computing with chaotic light.

Autor: Brückerhoff-Plückelmann F; Physical Institute, University of Münster, Münster, 48149, Germany.; Kirchhoff-Institute for Physics, University of Heidelberg, Heidelberg, 69120, Germany., Borras H; Institute of Computer Engineering, University of Heidelberg, Heidelberg, 69120, Germany., Klein B; Institute of Computer Engineering, University of Heidelberg, Heidelberg, 69120, Germany., Varri A; Physical Institute, University of Münster, Münster, 48149, Germany., Becker M; Institute for Geoinformatics, University of Münster, Münster, 48149, Germany.; Faculty of Mathematics & Computer Science, University of Münster, Münster, 48149, Germany., Dijkstra J; Kirchhoff-Institute for Physics, University of Heidelberg, Heidelberg, 69120, Germany., Brückerhoff M; DEVK RE, Cologne, 50668, Germany., Wright CD; Department of Engineering, University of Exeter, Exeter, EX44QF, UK., Salinga M; Institute of Materials Physics, University of Münster, Münster, 48149, Germany., Bhaskaran H; Department of Materials, University of Oxford, Oxford, OX43PJ, UK., Risse B; Institute for Geoinformatics, University of Münster, Münster, 48149, Germany.; Faculty of Mathematics & Computer Science, University of Münster, Münster, 48149, Germany., Fröning H; Institute of Computer Engineering, University of Heidelberg, Heidelberg, 69120, Germany., Pernice W; Physical Institute, University of Münster, Münster, 48149, Germany. wolfram.pernice@kip.uni-heidelberg.de.; Kirchhoff-Institute for Physics, University of Heidelberg, Heidelberg, 69120, Germany. wolfram.pernice@kip.uni-heidelberg.de.
Jazyk: angličtina
Zdroj: Nature communications [Nat Commun] 2024 Dec 01; Vol. 15 (1), pp. 10445. Date of Electronic Publication: 2024 Dec 01.
DOI: 10.1038/s41467-024-54931-6
Abstrakt: Biological neural networks effortlessly tackle complex computational problems and excel at predicting outcomes from noisy, incomplete data. Artificial neural networks (ANNs), inspired by these biological counterparts, have emerged as powerful tools for deciphering intricate data patterns and making predictions. However, conventional ANNs can be viewed as "point estimates" that do not capture the uncertainty of prediction, which is an inherently probabilistic process. In contrast, treating an ANN as a probabilistic model derived via Bayesian inference poses significant challenges for conventional deterministic computing architectures. Here, we use chaotic light in combination with incoherent photonic data processing to enable high-speed probabilistic computation and uncertainty quantification. We exploit the photonic probabilistic architecture to simultaneously perform image classification and uncertainty prediction via a Bayesian neural network. Our prototype demonstrates the seamless cointegration of a physical entropy source and a computational architecture that enables ultrafast probabilistic computation by parallel sampling.
Competing Interests: Competing interests: The authors declare no competing interests.
(© 2024. The Author(s).)
Databáze: MEDLINE
Externí odkaz: