Aggregation-aware MLP: An Unsupervised Approach for Graph Message-passing

TL;DR

This paper introduces AMLP, an unsupervised, adaptive MLP framework for graph message-passing that improves performance in heterophilic graphs without relying on labeled data.

Contribution

The paper presents AMLP, a novel unsupervised method that makes MLP adaptive to aggregation, addressing heterophily issues without requiring labeled data.

Findings

Outperforms existing methods on node clustering tasks.

Effective in heterophilic graph scenarios.

Demonstrates versatility across various graph learning tasks.

Abstract

Graph Neural Networks (GNNs) have become a dominant approach to learning graph representations, primarily because of their message-passing mechanisms. However, GNNs typically adopt a fixed aggregator function such as Mean, Max, or Sum without principled reasoning behind the selection. This rigidity, especially in the presence of heterophily, often leads to poor, problem dependent performance. Although some attempts address this by designing more sophisticated aggregation functions, these methods tend to rely heavily on labeled data, which is often scarce in real-world tasks. In this work, we propose a novel unsupervised framework, "Aggregation-aware Multilayer Perceptron" (AMLP), which shifts the paradigm from directly crafting aggregation functions to making MLP adaptive to aggregation. Our lightweight approach consists of two key steps: First, we utilize a graph reconstruction method that facilitates high-order grouping effects, and second, we employ a single-layer network to encode varying degrees of heterophily, thereby improving the capacity and applicability of the model. Extensive experiments on node clustering and classification demonstrate the superior performance of AMLP, highlighting its potential for diverse graph learning scenarios.