IJCAI 2025 Tutorial
Towards Low-Distortion Graph Representation Learning
The 34th International Joint Conference on Artificial Intelligence (IJCAI-25)

Tutorial description

Graph representation learning (GRL) has been extensively studied in both academia and industry. Essentially, it maps high-dimensional features and complex structures of graphs into a low-dimensional space for effective representation learning. However, a fundamental and critical problem is that the complex topology renders the low-dimensional mapping process highly sensitive to subtle changes in the data or the mapping process. Therefore, we suggest that GRL is prone to graph distortion, i.e., the loss of information regarding the graph's intrinsic complex structure and topological properties. Graph distortion manifests in three primary ways: noisy and perturbed structure, missing and tampered topological properties, and attribution fallacies in structural distribution. To address this significant issue, low-distortion GRL seeks to preserve the fidelity of intrinsic graph properties throughout the representation learning process. Low-distortion GRL has made great progress and attracted ever-increasing attention from the research community in recent years.

This tutorial aims to disseminate and promote the recent research achievement in low-distortion GRL, an exciting and fast-growing research direction within the broader fields of machine learning and artificial intelligence. We offer unique insights into low-distortion GRL from information theory, geometry, and causality. Specifically, we cover three major aspects, i.e., information-theoretic graph representation learning, geometry-guided graph representation learning, and invariance-guided graph representation learning. We will present novel, high-quality research findings, as well as innovative solutions to the challenging problems in low-distortion GRL and its applications. Besides, we will discuss future directions, such as advanced theories, low-distortion approaches in graph large language models and graph foundation model, and novel applications in AI for Science.