Abstract:The dreaming Hopfield model constitutes a generalization of the Hebbian paradigm for neural networks, that is able to perform on-line learning when "awake" and also to account for off-line "sleeping" mechanisms. The latter have been shown to enhance storing in such a way that, in the long sleep-time limit, this model can reach the maximal storage capacity achievable by networks equipped with symmetric pairwise interactions. In this paper, we inspect the minimal amount of information that must be supplied to such a network to guarantee a successful generalization, and we test it both on random synthetic and on standard structured datasets (i.e., MNIST, Fashion-MNIST and Olivetti). By comparing these minimal thresholds of information with those required by the standard (i.e., always "awake") Hopfield model, we prove that the present network can save up to ∼90% of the dataset size, yet preserving the same performance of the standard counterpart. This suggests that sleep may play a pivotal role in explaining the gap between the large volumes of data required to train artificial neural networks and the relatively small volumes needed by their biological counterparts. Further, we prove that the model Cost function (typically used in statistical mechanics) admits a representation in terms of a standard Loss function (typically used in machine learning) and this allows us to analyze its emergent computational skills both theoretically and computationally: a quantitative picture of its capabilities as a function of its control parameters is achieved and consistency between the two approaches is highlighted. The resulting network is an associative memory for pattern recognition tasks that learns from examples on-line, generalizes correctly (in suitable regions of its control parameters) and optimizes its storage capacity by off-line sleeping: such a reduction of the training cost can be inspiring toward sustainable AI and in situations where data are relatively sparse.

Daydreaming Hopfield Networks and their surprising effectiveness on correlated data

Hebbian dreaming for small datasets

Dreaming neural networks: Forgetting spurious memories and reinforcing pure ones

Effect of spatial correlations on Hopfield Neural Network and Dense Associative Memories

Eigenvector Dreaming

A generalized Hopfield model to store and retrieve mismatched memory patterns

Dreaming Learning

Regularization, early-stopping and dreaming: a Hopfield-like setup to address generalization and overfitting

Memory Dynamics in Attractor Networks with Saliency Weights.

Interpolating between boolean and extremely high noisy patterns through Minimal Dense Associative Memories

Learning, Storing, and Disentangling Correlated Patterns in Neural Networks.

Uniform Memory Retrieval with Larger Capacity for Modern Hopfield Models

Random Features Hopfield Networks generalize retrieval to previously unseen examples

Storage and Learning phase transitions in the Random-Features Hopfield Model

Interacting dreaming neural networks

High storage capacity in the Hopfield model with auto-interactions - stability analysis

Input-Driven Dynamics for Robust Memory Retrieval in Hopfield Networks

Dense Associative Memory Through the Lens of Random Features

A Feed-Forward Neural Network for Increasing the Hopfield-Network Storage Capacity