If Neural Networks Are Allowed To Sleep And Dream, Their Performance Sensibly Increases

The harmonic oscillator for associative memory and pattern recognition in Artificial Intelligence is certainly the Hopfield model [1] (or, equivalently [2], its dual representation, i.e. the Restricted Boltzmann Machine (RBM) [3]). In a nutshell, we can store information (consisting in a set of P digital words or -generally speaking- patterns of information) by suitably modifying the synaptic interactions among neurons in the Hopfield neural network by means of the so-called Hebbian learning (or by using contrastive divergence algorithms for training RBMs).

However, since the seminal AGS theory [5], we know that -for this model- the critical capacity αc (namely the maximal amount of patterns Pmax a network equipped with N neurons can handle, i.e. αc = Pmax/N), is αc ∼ 0.14. On the other side, general information theoretical considerations (i.e., Gardner Theory [4]) prescribe that the maximal critical capacity for networks equipped with symmetric couplings (as the Hopfield model) is precisely αc = 1 (i.e. Pmax = N), significantly better than the Hopfield case.


Recently, in our work [6], we have shown that forcing the network to sleep -namely, mathematically implementing dreaming processes (mimicking slow wave sleep (SWS) and random eye movements (REM) phases in real mammal’s brain), after a proper rest the Hopfield model has impressive enhanced skills. In particular, it reaches the maximal bound for the critical storage, i.e. αc = 1, further, stored patterns are very stable configurations (technically speaking, they are free energy absolute minima) up to α ∼ 0.85 (while in standard Hopfield picture, this is true solely for α < 0.05). This is a remarkable property as, in the long time limit, any stochastic neural dynamics for the standard Hopfield model, if α > 0.05, almost surely ends up in a (spin-glass) state that does not overlap significantly with any of the stored patterns; while -if the network can sleep- this happens only for α > 0.85 conferring to network’s functionalities an enhanced robustness with respect to severe loads of patterns.

Finally, thanks to the equivalence between RBMs and Hopfield networks [2], we know that the ratio α = P/N in the Hopfield retrieval matches the ratio among the size of the hidden layer (built of P hidden neurons) over that of the visible one (build of N visible neurons) for RBMs. In the standard scenario, if the hidden layer is larger than ∼ 0.14 times the visible one, learning can be prone to over-fitting (as its dual Hopfield model would lie already outside the working regime, in the spin-glass phase). Allowing the network to sleep, we can learn without over-fitting even by keeping the two layers roughly of the same size (and this obviously allows to extract much more features from the data-sets presented to the network during the training stage).

Summarizing, we believe that in the process of cognition -while certainly learning and retrieval keep covering a pivotal role – also sleeping is mandatory for Artificial Intelligence, as it is for the biological one.

These findings are described in the article entitled Dreaming neural networks: Forgetting spurious memories and reinforcing pure ones, recently published in the journal Neural Networks.



  1. J.J. Hopfield, Neural networks and physical systems with emergent collective computational abilities, Proc. Natl. Acad. Sci. 79(8):2554, (1982).
  2. A. Barra, et al., On the equivalence of Hopfield networks and Boltzmann machines, Neural Networks 34:1, (2012).
  3. D.H. Ackley, G.E. Hinton, T.J. Sejnowski, A learning algorithm for Boltzmann machines, Cognitive science 9(1):147, (1985).
  4. E. Gardner, The space of interactions in neural network models, J. Phys. A 21(1):257, (1998).
  5. D.J. Amit, H. Gutfreund, H. Sompolinsky, Storing infinite numbers of patterns in a spin-glass model of neural networks, Phys. Rev. Lett. 55(14):1530, (1985).
  6. A. Fachechi, E. Agliari, A. Barra, Dreaming neural networks: forgetting spurious memories and reinforcing pure ones, Neural Networks (2019) in press



Strategies For Encouraging Solar Technology Adoption Through Public Engagement And Visual Exposure

Globally, electricity systems are undergoing a significant transformation to green energy in response to technology change and climate policy. Yet, […]

What Does The Future Hold For The Endangered Cantabrian Brown Bear?

Brown bears, like many other large carnivores, are recovering and expanding their range in recent years, especially in Europe. In […]

“Psychobiotics” And The Science Of How Gut Bacteria Can Affect The Human Brain

The connection between the digestive system and the brain is not something new to scientists; the term “gut-brain axis” has […]

Can We Use Entomopathogenic Fungi As Endophytes For Biological Control Of The Horse-Chestnut Leaf Miner?

The horse-chestnut leaf miner, Cameraria ohridella (Lepidoptera, Gracillariidae), is a major pest of horse-chestnut trees (Aesculus hippocastanum) in Europe. This […]

Borax For Ants: A Simple And Effective Ant Killer

Need a quick and easy way to kill ants, use this Borax recipe! You should mix a half cup of […]

How Large Meteorite Impacts Affect Planetary Mantle Convection

Every now and then, car- or house-sized bodies entering Earth’s atmosphere from outer space meet a fiery end in a […]

Map Of The Amalfi Coast In Italy

The Amalfi coast map shows you a serene, picturesque stretch of coastline found in Italy on the northern coast of […]

Science Trends is a popular source of science news and education around the world. We cover everything from solar power cell technology to climate change to cancer research. We help hundreds of thousands of people every month learn about the world we live in and the latest scientific breakthroughs. Want to know more?