- Physical Review Journals
All JournalsPhysics Magazine

Physical Review Research

Reuse & Permissions

It is not necessary to obtain permission to reuse this article or its components as it is available under the terms of the Creative Commons Attribution 4.0 International license. This license permits unrestricted use, distribution, and reproduction in any medium, provided attribution to the author(s) and the published article's title, journal citation, and DOI are maintained. Please note that some figures may have been included with permission from other third parties. It is your responsibility to obtain the proper permission from the rights holder directly for these figures.

Export citation

Export citation

Choose format for download:

Download Citation
  • Letter
  • Open Access

Neuromodulation via Krotov-Hopfield improves accuracy and robustness of restricted Boltzmann machines

Başer Tambaş* and A. Levent Subaşı†

Alkan Kabakçıoğlu‡

  • *Contact author: tambas19@itu.edu.tr
  • †Contact author: alsubasi@itu.edu.tr
  • ‡Contact author: akabakcioglu@ku.edu.tr
PDF

Phys. Rev. Research 8, L012004 – Published 7 January, 2026

DOI: https://doi.org/10.1103/y82t-c34b

Abstract

In biological systems, neuromodulation tunes synaptic plasticity based on the internal state of the organism, complementing stimulus-driven Hebbian learning. The algorithm recently proposed by Krotov and Hopfield (KH) [D. Krotov and J. J. Hopfield, Proc. Natl. Acad. Sci. USA 116, 7723 (2019)] can be utilized to mirror this process in artificial neural networks, where its built-in intralayer competition and selective inhibition of synaptic updates offer a cost-effective remedy for the lack of lateral connections through a simplified attention mechanism. We demonstrate that KH-modulated restricted Boltzmann machines (RBMs) outperform standard (shallow) RBMs in both reconstruction and classification tasks, offering a superior trade-off between generalization performance and model size, with the additional benefit of robustness to weight initialization as well as to overfitting during training.

View figure in article

Physics Subject Headings (PhySH)

Article Text

Supplemental Material

References (41)

  1. S. Schmidgall, R. Ziaei, J. Achterberg, L. Kirsch, S. P. Hajiseyedrazi, and J. Eshraghian, Brain-inspired learning in artificial neural networks: A review, APL Mach. Learn. 2, 021501 (2024).
  2. Y. LeCun, Y. Bengio, and G. Hinton, Deep learning, Nature (London) 521, 436 (2015).
  3. D. Rumelhart, G. Hinton, and R. Williams, Learning representations by back-propagating errors, Nature (London) 323, 533 (1986).
  4. T. P. Lillicrap, D. Cownden, D. B. Tweed, and C. J. Akerman, Random synaptic feedback weights support error backpropagation for deep learning, Nat. Commun. 7, 13276 (2016).
  5. D.-H. Lee, S. Zhang, A. Fischer, and Y. Bengio, Difference target propagation, in Machine Learning and Knowledge Discovery in Databases, edited by A. Appice, P. P. Rodrigues, V. Santos Costa, C. Soares, J. Gama, and A. Jorge (Springer International Publishing, Cham, 2015), pp. 498–515.
  6. B. Scellier and Y. Bengio, Equilibrium propagation: Bridging the gap between energy-based models and backpropagation, Front. Comput. Neurosci. 11, 24 (2017).
  7. D. H. Ackley, G. E. Hinton, and T. J. Sejnowski, A learning algorithm for Boltzmann machines, Cognit. Sci. 9, 147 (1985).
  8. G. E. Hinton, Training products of experts by minimizing contrastive divergence, Neural Comput. 14, 1771 (2002).
  9. D. Krotov and J. J. Hopfield, Unsupervised learning by competing hidden units, Proc. Natl. Acad. Sci. USA 116, 7723 (2019).
  10. L. Squadrani, N. Curti, E. Giampieri, D. Remondini, B. Blais, and G. Castellani, Effectiveness of biologically inspired neural network models in learning and pattern memorization, Entropy 24, 682 (2022).
  11. E. Bienenstock, L. Cooper, and P. Munro, Theory for the development of neuron selectivity: Orientation specificity and binocular interaction in visual cortex, J. Neurosci. 2, 32 (1982).
  12. D. O. Hebb, The Organization of Behavior: A Neuropsychological Theory (John Wiley and Sons, New York, 1949).
  13. R. Salakhutdinov and G. E. Hinton, Deep Boltzmann machines, in Proceedings of the 12th International Conference on Artificial Intelligence and Statistics (AISTATS), edited by D. van Dyk and M. Welling (Machine Learning Research, Florida, 2009), Vol. 5, pp. 448–455.
  14. M. Mézard and A. Montanari, Information, Physics, and Computation (Oxford University Press, Oxford, 2009).
  15. A. Decelle, G. Fissore, and C. Furtlehner, Thermodynamics of restricted Boltzmann machines and related learning dynamics, J. Stat. Phys. 172, 1576 (2018).
  16. E. W. Tramel, M. Gabrié, A. Manoel, F. Caltagirone, and F. Krzakala, Deterministic and generalized framework for unsupervised learning with restricted Boltzmann machines, Phys. Rev. X 8, 041006 (2018).
  17. E. Agliari, A. Barra, and B. Tirozzi, Free energies of Boltzmann machines: Self-averaging, annealed and replica symmetric approximations in the thermodynamic limit, J. Stat. Mech. (2019) 033301.
  18. A. Decelle and C. Furtlehner, Restricted Boltzmann machine: Recent advances and mean-field theory, Chin. Phys. B 30, 040202 (2021).
  19. A. Fachechi, E. Agliari, M. Aquaro, A. Coolen, and M. Mulder, Fundamental operating regimes, hyper-parameter fine-tuning and glassiness: Towards an interpretable replica-theory for trained restricted Boltzmann machines, J. Phys. A: Math. Theor. 58, 065004 (2025).
  20. H. Lee, R. Grosse, R. Ranganath, and A. Y. Ng, Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations, in Proceedings of the 26th Annual International Conference on Machine Learning (ICML) (Association for Computing Machinery, Montreal, Canada, 2009), pp. 609–616.
  21. G. Shi, J. Zhang, N. Ji, and C. Wang, A new variant of restricted Boltzmann machine with horizontal connections, Neural Comput. Appl. 31, 6521 (2019).
  22. F. Nadim and D. Bucher, Neuromodulation of neurons and synapses, Curr. Opin. Neurobiol. 29, 48 (2014).
  23. J. M. Shine, Neuromodulatory influences on integration and segregation in the brain, Trends Cognit. Sci. 23, 572 (2019).
  24. N. Fremaux and W. Gerstner, Neuromodulated spike-timing-dependent plasticity, and theory of three-factor learning rules, Front. Neural Circuits 9, 85 (2016).
  25. J. Mei, R. Meshkinnejad, and Y. Mohsenzadeh, Effects of neuromodulation-inspired mechanisms on the performance of deep neural networks in a spatial learning task, iScience 26, 106026 (2023).
  26. L. Grinberg, J. Hopfield, and D. Krotov, Local unsupervised learning for image analysis, arXiv:1908.08993.
  27. S. Osindero and G. Hinton, Modeling image patches with a directed hierarchy of Markov random fields, in Proceedings of the 21st International Conference on Neural Information Processing Systems, NIPS’07 (Curran Associates Inc., Red Hook, NY, 2007), pp. 1121–1128.
  28. See Supplemental Material at http://link.aps.org/supplemental/10.1103/y82t-c34b for details on hyperparameter selection, TD/BU-RBM comparison, and experiments on scheduling and KMNIST (a more challenging learning task), which includes Refs.  [35, 37, 39].
  29. A. Fischer and C. Igel, An introduction to restricted Boltzmann machines, in Progress in Pattern Recognition, Image Analysis, Computer Vision, and Application Lecture Notes in Computer Science, edited by L. Alvarez, M. Mejail, L. Gomez, and J. Jacobo (Springer, Berlin, 2012), pp. 14–36.
  30. T. Tieleman, Training restricted Boltzmann machines using approximations to the likelihood gradient, in Proceedings of the 25th International Conference on Machine Learning (ICML’08) (ACM Press, Helsinki, Finland, 2008), pp. 1064–1071.
  31. M. Welling and G. E. Hinton, A new learning algorithm for mean-field Boltzmann machines, in Artificial Neural Networks—ICANN 2002, edited by G. Goos, J. Hartmanis, J. Van Leeuwen, and J. R. Dorronsoro (Springer, Berlin, 2002), Vol. 2415, pp. 351–357.
  32. M. Gabrie, E. W. Tramel, and F. Krzakala, Training restricted Boltzmann machine via the Thouless-Anderson-Palmer free energy, in Advances in Neural Information Processing Systems, edited by C. Cortes, N. Lawrence, D. Lee, M. Sugiyama, and R. Garnett (Curran Associates, Inc., Montreal, Canada, 2015), Vol. 28.
  33. G. E. Hinton, A practical guide for training restricted Boltzmann machines, in Neural Networks: Tricks of the Trade, 2nd ed., Lecture Notes in Computer Science, edited by G. Montavon, G. B. Orr, and K.-R. Müller (Springer, Berlin, 2012), pp. 599–619.
  34. D. Krotov, Biological learning, GitHub repository, 2020, https://github.com/DimaKrotov/Biological_Learning.
  35. Y. LeCun, C. Cortes, and C. Burges, MNIST handwritten digit database (1998), http://yann.lecun.com/exdb/mnist/.
  36. Y. A. LeCun, L. Bottou, G. B. Orr, and K.-R. Müller, Efficient backprop, in Neural Networks: Tricks of the Trade: Second Edition, edited by G. Montavon, G. B. Orr, and K.-R. Müller (Springer, Berlin, 2012), pp. 9–48.
  37. The code implementations are made available on GitHub: https://github.com/basertambas/KH-RBM.
  38. H. Larochelle, M. Mandel, R. Pascanu, and Y. Bengio, Learning algorithms for the classification restricted Boltzmann machine, J. Mach. Learn. Res. 13, 643 (2012).
  39. T. Clanuwat, M. Bober-Irizar, A. Kitamoto, A. Lamb, K. Yamamoto, and D. Ha, Deep learning for classical Japanese literature, https://doi.org/10.20676/00000341.
  40. G. E. Hinton, N. Srivastava, A. Krizhevsky, I. Sutskever, and R. R. Salakhutdinov, Improving neural networks by preventing co-adaptation of feature detectors, arXiv:1207.0580.
  41. M. Welling and Y. W. Teh, Bayesian learning via stochastic gradient Langevin dynamics, in Proceedings of the 28th International Conference on Machine Learning (ICML-11) (Omnipress, Washington, 2011), pp. 681–688.

Outline

Information

Phys. Rev. Research 8, L012004– Published 7 January, 2026
  • Received 10 April 2025
  • Revised 7 August 2025
  • Accepted 5 December 2025
Crossmark - Check for updates button
Creative Commons Logo

Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article's title, journal citation, and DOI.

Published by the American Physical Society

Sign In to Your Journals Account

Filter

Filter

Article Lookup

Enter a citation