Social Hippocampus Memory Learning

TL;DR

SoHip introduces a memory-sharing approach for heterogeneous federated learning, inspired by hippocampus mechanisms, enabling privacy-preserving collaboration that improves prediction accuracy without sharing raw data or models.

Contribution

This work presents a novel memory-centric framework for social machine learning that enhances collaboration among heterogeneous agents via hippocampus-inspired memory sharing.

Findings

Outperforms existing methods with up to 8.78% accuracy improvement.

Provides theoretical analysis on convergence and privacy.

Effective in benchmark datasets with multiple baselines.

Abstract

Social learning highlights that learning agents improve not in isolation, but through interaction and structured knowledge exchange with others. When introduced into machine learning, this principle gives rise to social machine learning (SML), where multiple agents collaboratively learn by sharing abstracted knowledge. Federated learning (FL) provides a natural collaboration substrate for this paradigm, yet existing heterogeneous FL approaches often rely on sharing model parameters or intermediate representations, which may expose sensitive information and incur additional overhead. In this work, we propose SoHip (Social Hippocampus Memory Learning), a memory-centric social machine learning framework that enables collaboration among heterogeneous agents via memory sharing rather than model sharing. SoHip abstracts each agent's individual short-term memory from local representations, consolidates it into individual long-term memory through a hippocampus-inspired mechanism, and fuses it with collectively aggregated long-term memory to enhance local prediction. Throughout the process, raw data and local models remain on-device, while only lightweight memory are exchanged. We provide theoretical analysis on convergence and privacy preservation properties. Experiments on two benchmark datasets with seven baselines demonstrate that SoHip consistently outperforms existing methods, achieving up to 8.78% accuracy improvements.