Self-attention as an attractor network: transient memories without backpropagation
Abstract
Transformers are one of the most successful architectures of modern neural networks. At their core there is the so-called attention mechanism, which recently interested the physics community as it can be written as the derivative of an energy function in certain cases: while it is possible to write the cross-attention layer as a modern Hopfield network, the same is not possible for the self-attention, which is used in the GPT architectures and other autoregressive models. In this work we show that it is possible to obtain the self-attention layer as the derivative of local energy terms, which resemble a pseudo-likelihood. We leverage the analogy with pseudo-likelihood to design a recurrent model that can be trained without backpropagation: the dynamics shows transient states that are strongly correlated with both train and test examples. Overall we present a novel framework to interpret self-attention as an attractor network, potentially paving the way for new theoretical approaches inspired from physics to understand transformers.
Keywords
Cite
@article{arxiv.2409.16112,
title = {Self-attention as an attractor network: transient memories without backpropagation},
author = {Francesco D'Amico and Matteo Negri},
journal= {arXiv preprint arXiv:2409.16112},
year = {2024}
}