DK-Root: A Joint Data-and-Knowledge-Driven Framework for Root Cause Analysis of QoE Degradations in Mobile Networks

TL;DR

DK-Root addresses root-cause analysis for QoE degradations in mobile networks by fusing abundant rule-based weak supervision with scarce expert annotations. It introduces a three-stage pipeline: conditional diffusion-based KPI augmentation, diffusion-guided supervised contrastive representation learning, and expert-guided fine-tuning of a lightweight classifier. The method treats KPI time-series $X$ as an $m×l$ real-valued object and leverages dual supervision sources $(D_r, D_e)$ to learn robust representations and accurate RCA decisions. Experiments on operator-grade data show state-of-the-art accuracy and ablations validating the necessity of conditional diffusion augmentation and the pretrain-finetune design, highlighting practical impact for scalable mobile-network RCA.

Abstract

Diagnosing the root causes of Quality of Experience (QoE) degradations in operational mobile networks is challenging due to complex cross-layer interactions among kernel performance indicators (KPIs) and the scarcity of reliable expert annotations. Although rule-based heuristics can generate labels at scale, they are noisy and coarse-grained, limiting the accuracy of purely data-driven approaches. To address this, we propose DK-Root, a joint data-and-knowledge-driven framework that unifies scalable weak supervision with precise expert guidance for robust root-cause analysis. DK-Root first pretrains an encoder via contrastive representation learning using abundant rule-based labels while explicitly denoising their noise through a supervised contrastive objective. To supply task-faithful data augmentation, we introduce a class-conditional diffusion model that generates KPIs sequences preserving root-cause semantics, and by controlling reverse diffusion steps, it produces weak and strong augmentations that improve intra-class compactness and inter-class separability. Finally, the encoder and the lightweight classifier are jointly fine-tuned with scarce expert-verified labels to sharpen decision boundaries. Extensive experiments on a real-world, operator-grade dataset demonstrate state-of-the-art accuracy, with DK-Root surpassing traditional ML and recent semi-supervised time-series methods. Ablations confirm the necessity of the conditional diffusion augmentation and the pretrain-finetune design, validating both representation quality and classification gains.

Figures (9)

• Figure 1: Poor Communication Quality of Experience Caused by Different Root-causes.
• Figure 2: Overview of the root-cause analysis task for QoE degradations in mobile networks.
• Figure 3: The flowchart of the proposed DK-Root framework.
• Figure 4: The workflow of the proposed conditional diffusion model.
• Figure 5: The flowchart of the Stage II.