Emergent Self-Attention from Astrocyte-Gated Associative Memory Dynamics

TL;DR

This paper presents a novel associative memory model where astrocyte-like modulation enables emergent self-attention, improving retrieval accuracy and linking glial dynamics to attention mechanisms.

Contribution

It introduces a Hopfield-type model with astrocytic modulation, demonstrating global convergence and emergent self-attention through a dynamical systems framework.

Findings

Model significantly improves retrieval accuracy under high load.

Astrocytic gains implement a softmax-normalized allocation over pattern similarity.

Framework links glial modulation to attention-like computation.

Abstract

We introduce a Hopfield-type associative memory in which effective connectivity is multiplicatively modulated by astrocytic gains evolving under an entropy-regularized replicator equation. The coupled neuron-astrocyte dynamics admit a Lyapunov function, ensuring global convergence. At fixed points, astrocytic gains implement a softmax-normalized allocation over pattern similarity scores, yielding a mechanistic realization of self-attention as emergent routing on the gain simplex. In regimes of high memory load and interference, the model significantly improves retrieval accuracy relative to classical Hopfield dynamics and recent neuron-astrocyte baselines. These results establish a dynamical systems framework linking glial modulation, competitive resource allocation, and attention-like computation.