Efficient Graph Optimization via Distance-Aware Graph Representation Learning

TL;DR

This paper introduces DRTR, a novel graph neural network framework that enhances structural dependency capture through distance-aware message passing and dynamic topology refinement, improving accuracy and scalability.

Contribution

It presents a new method combining static preprocessing and dynamic resampling for better graph representation learning, outperforming standard GNNs.

Findings

DRTR achieves higher accuracy than baseline GNNs.

DRTR maintains scalability with only 20% additional computation.

Effective in complex and noisy graph environments.

Abstract

We propose an efficient framework that integrates distance-aware multi-hop message passing with dynamic topology refinement. Unlike standard GNNs that rely on shallow, fixed-hop aggregation, DRTR leverages both static preprocessing and dynamic resampling to capture deeper structural dependencies. A \emph{Distance Recomputator} prunes semantically weak edges using adaptive attention, while a \emph{Topology Reconstructor} establishes latent connections among distant but relevant nodes. This joint mechanism enables more expressive and robust graph representation optimization across evolving graph structures. Extensive experiments demonstrate that DRTR outperforms baseline GNNs in both accuracy and scalability, with at most 20\% computational overhead, especially in complex and noisy graph environments.