GrAMME: Semi-Supervised Learning using Multi-layered Graph Attention Models

TL;DR

This paper introduces GrAMME, a semi-supervised learning framework for multi-layered graphs that leverages attention models to improve node classification, demonstrating significant performance gains over existing methods.

Contribution

The paper proposes two novel attention-based architectures, GrAMME-SG and GrAMME-Fusion, for effective feature learning in multi-layered graphs, addressing a gap in current graph neural network approaches.

Findings

Significant performance improvements over state-of-the-art methods.

Random node features can be effective even without explicit attributes.

Effective modeling of inter-layer dependencies enhances node classification.

Abstract

Modern data analysis pipelines are becoming increasingly complex due to the presence of multi-view information sources. While graphs are effective in modeling complex relationships, in many scenarios a single graph is rarely sufficient to succinctly represent all interactions, and hence multi-layered graphs have become popular. Though this leads to richer representations, extending solutions from the single-graph case is not straightforward. Consequently, there is a strong need for novel solutions to solve classical problems, such as node classification, in the multi-layered case. In this paper, we consider the problem of semi-supervised learning with multi-layered graphs. Though deep network embeddings, e.g. DeepWalk, are widely adopted for community discovery, we argue that feature learning with random node attributes, using graph neural networks, can be more effective. To this end, we propose to use attention models for effective feature learning, and develop two novel architectures, GrAMME-SG and GrAMME-Fusion, that exploit the inter-layer dependencies for building multi-layered graph embeddings. Using empirical studies on several benchmark datasets, we evaluate the proposed approaches and demonstrate significant performance improvements in comparison to state-of-the-art network embedding strategies. The results also show that using simple random features is an effective choice, even in cases where explicit node attributes are not available.