MILD: Multi-Intent Learning and Disambiguation for Proactive Failure Prediction in Intent-based Networking

2026-02-15

AI Summary

This study addresses the challenge of identifying root-cause intents in multi-intent driven networks, where single-point failures often trigger correlated KPI drifts across multiple intents, obscuring causal attribution. To overcome this, the work proposes a paradigm shift from passive detection to proactive prediction, enabling fault anticipation and root-cause localization within a fixed time window. The authors design a teacher-enhanced Mixture-of-Experts (MoE) architecture featuring a gating disambiguation module, intent-specific prediction heads, and an interpretable KPI mechanism to effectively decouple co-drifting signals and calibrate risk. Experimental results demonstrate that the proposed approach improves root-cause intent identification accuracy by 9.4%–45.8% and advances fault remediation lead time by 3.8%–92.5% under nonlinear failure and co-drift scenarios.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Reasoning under Uncertainty: Causality

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

In multi-intent intent-based networks, a single fault can trigger co-drift where multiple intents exhibit symptomatic KPI degradation, creating ambiguity about the true root-cause intent. We present MILD, a proactive framework that reformulates intent assurance from reactive drift detection to fixed-horizon failure prediction with intent-level disambiguation. MILD uses a teacher-augmented Mixture-of-Experts where a gated disambiguation module identifies the root-cause intent while per-intent heads output calibrated risk scores. On a benchmark with non-linear failures and co-drifts, MILD provides 3.8\%--92.5\% longer remediation lead time and improves intent-level root-cause disambiguation accuracy by 9.4\%--45.8\% over baselines. MILD also provides per-alert KPI explanations, enabling actionable diagnosis.

Problem

• intent-based networking
• multi-intent
• co-drift
• root-cause disambiguation
• failure prediction

Innovation

• intent-based networking
• multi-intent disambiguation
• proactive failure prediction
• Mixture-of-Experts
• root-cause analysis