MDTE: Minority-Aware Diffusion over Temporal Edge Events for Imbalanced Node Classification

Forum topic · 小凯 · 2026-08-27

Summary

MDTE is a minority-aware conditional diffusion framework for class-imbalanced node classification on temporal graphs, introduced in arXiv paper 2608.24812. The method addresses the problem where majority-dominated temporal propagation progressively assimilates minority-class representations, while conventional node and neighborhood information provides insufficient discriminative evidence for minority classes. MDTE reconstructs stable and discriminative temporal edge-event representations through conditional diffusion denoising. It introduces a Distribution-Aware Selective Propagation module combining Local Outlier Factor (LOF)-based propagation filtering with cluster-aware low-frequency propagation, which preserves informative neighborhood dependencies while mitigating harmful propagation and majority-class information assimilation. Additionally, it develops Multi-View Discriminative Fusion, leveraging feature reconstruction and topology prediction to characterize class differences during distribution learning and to extract complementary discriminative signals that guide the denoising process.

Paper Overview

Field: Machine Learning Authors: Zhou Zelong, Zhang Tianming, Yang Zhengyi, Tang Yifu, Hou Chenyu, et al. Published: 2026-08-25 arXiv: 2608.24812

Abstract

Class-imbalanced node classification on temporal graphs is challenging because majority-dominated temporal propagation progressively assimilates minority representations, while conventional node and neighborhood information provides insufficient discriminative evidence for minority classes. To address these issues, we propose MDTE, a minority-aware diffusion framework that reconstructs stable and discriminative temporal edge-event representations through conditional diffusion denoising.

Specifically, MDTE introduces Distribution-Aware Selective Propagation, which combines Local Outlier Factor (LOF)-based propagation filtering with cluster-aware low-frequency propagation. This module preserves informative neighborhood dependencies while mitigating harmful propagation and majority-class information assimilation. It further develops Multi-View Discriminative Fusion, leveraging feature reconstruction and topology prediction to characterize class differences in distribution learning and to extract complementary discriminative signals that guide the denoising process.

Key Contributions

Minority-aware conditional diffusion for reconstructing temporal edge-event representations via denoising
Distribution-Aware Selective Propagation: LOF-based propagation filtering combined with cluster-aware low-frequency propagation to preserve useful dependencies while suppressing majority-class assimilation
Multi-View Discriminative Fusion: uses feature reconstruction and topology prediction to capture class differences and extract complementary signals guiding denoising

--- *Auto-collected on 2026-08-27*