Hyperdimensional Representation Learning for Node Classification and Link Prediction

2024-02-26

AI Summary

This work addresses node classification and link prediction on graph-structured data. We propose Hyperdimensional Graph Learning (HDGL), a novel framework that maps node features into a high-dimensional binary space and aggregates neighborhood information in a single forward pass via hyperdimensional operations—specifically bundling and binding—enabling joint modeling of both tasks. HDGL is the first to systematically integrate hyperdimensional computing into graph representation learning, eliminating reliance on multi-layer iterative propagation, gradient-based optimization, and extensive hyperparameter tuning inherent in conventional GNNs. Theoretically, HDGL guarantees injectivity of node representations and compatibility with downstream tasks. Experiments demonstrate that HDGL achieves node classification accuracy competitive with state-of-the-art GNNs while reducing inference cost significantly; for link prediction, it matches DeepWalk’s performance, offering both efficiency and broad applicability across diverse graph domains.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Humans and AI: Human-in-the-loop Machine Learning
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

We introduce Hyperdimensional Graph Learner (HDGL), a novel method for node classification and link prediction in graphs. HDGL maps node features into a very high-dimensional space ( extit{hyperdimensional} or HD space for short) using the emph{injectivity} property of node representations in a family of Graph Neural Networks (GNNs) and then uses HD operators such as extit{bundling} and extit{binding} to aggregate information from the local neighborhood of each node yielding latent node representations that can support both node classification and link prediction tasks. HDGL, unlike GNNs that rely on computationally expensive iterative optimization and hyperparameter tuning, requires only a single pass through the data set. We report results of experiments using widely used benchmark datasets which demonstrate that, on the node classification task, HDGL achieves accuracy that is competitive with that of the state-of-the-art GNN methods at substantially reduced computational cost; and on the link prediction task, HDGL matches the performance of DeepWalk and related methods, although it falls short of computationally demanding state-of-the-art GNNs.

Problem

• Develops HDGL for graph analysis
• Improves computational efficiency in GNNs
• Supports node classification and link prediction

Innovation

• Hyperdimensional space mapping
• Single-pass data processing
• Competitive accuracy with reduced cost