Abstract:We propose a two-stage memory retrieval dynamics for modern Hopfield models, termed $\mathtt{U\text{-}Hop}$, with enhanced memory capacity. Our key contribution is a learnable feature map $\Phi$ which transforms the Hopfield energy function into kernel space. This transformation ensures convergence between the local minima of energy and the fixed points of retrieval dynamics within the kernel space. Consequently, the kernel norm induced by $\Phi$ serves as a novel similarity measure. It utilizes the stored memory patterns as learning data to enhance memory capacity across all modern Hopfield models. Specifically, we accomplish this by constructing a separation loss $\mathcal{L}_\Phi$ that separates the local minima of kernelized energy by separating stored memory patterns in kernel space. Methodologically, $\mathtt{U\text{-}Hop}$ memory retrieval process consists of: (Stage I) minimizing separation loss for a more uniform memory (local minimum) distribution, followed by (Stage II) standard Hopfield energy minimization for memory retrieval. This results in a significant reduction of possible metastable states in the Hopfield energy function, thus enhancing memory capacity by preventing memory confusion. Empirically, with real-world datasets, we demonstrate that $\mathtt{U\text{-}Hop}$ outperforms all existing modern Hopfield models and state-of-the-art similarity measures, achieving substantial improvements in both associative memory retrieval and deep learning tasks. Code is available at <a class="link-external link-https" href="https://github.com/MAGICS-LAB/UHop" rel="external noopener nofollow">this https URL</a> ; future updates are on <a class="link-https" data-arxiv-id="2404.03827" href="https://arxiv.org/abs/2404.03827">arXiv:2404.03827</a>

Dynamic Capacity Estimation in Hopfield Networks

Storage Capacity of the Hopfield Network Associative Memory

Capacity of the Hebbian-Hopfield network associative memory

The capacity of the Hopfield associative memory

Uniform Memory Retrieval with Larger Capacity for Modern Hopfield Models

Simplicial Hopfield networks

On Computational Limits of Modern Hopfield Models: A Fine-Grained Complexity Analysis

Input-Driven Dynamics for Robust Memory Retrieval in Hopfield Networks

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

Estimation on Domain of Attraction and Convergence Rate of Hopfield Continuous Feedback Neural Networks

A theoretical investigation into the performance of the Hopfield model

Information capacity of a network of spiking neurons

Robustness in Hopfield neural networks with biased memory patterns

Maximizing memory capacity in heterogeneous networks

Exponential Capacity of Dense Associative Memories

Long Sequence Hopfield Memory

Estimation of the domain of attraction and the convergence rate of a Hopfield associative memory and an application

Provably Optimal Memory Capacity for Modern Hopfield Models: Transformer-Compatible Dense Associative Memories as Spherical Codes

Nonparametric Modern Hopfield Models

Long sequence Hopfield memory*

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