![[2602.14283] MILD: Multi-Intent Learning and Disambiguation for Proactive Failure Prediction in Intent-based Networking](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.14283] MILD: Multi-Intent Learning and Disambiguation for Proactive Failure Prediction in Intent-based Networking
Summary
The paper presents MILD, a proactive framework for failure prediction in intent-based networking, enhancing root-cause intent disambiguation and remediation lead time.
Why It Matters
As networks become increasingly complex, proactive failure prediction is critical for maintaining performance. MILD addresses the ambiguity in intent-based networks, offering a solution that improves response times and accuracy in identifying root causes, which is essential for network reliability and efficiency.
Key Takeaways
- MILD reformulates intent assurance from reactive to proactive failure prediction.
- The framework improves remediation lead time by 3.8% to 92.5%.
- Intent-level disambiguation accuracy is enhanced by 9.4% to 45.8% over existing methods.
- MILD provides actionable KPI explanations for better diagnosis.
- The approach utilizes a teacher-augmented Mixture-of-Experts model.
Computer Science > Networking and Internet Architecture arXiv:2602.14283 (cs) [Submitted on 15 Feb 2026] Title:MILD: Multi-Intent Learning and Disambiguation for Proactive Failure Prediction in Intent-based Networking Authors:Md. Kamrul Hossain, Walid Aljoby View a PDF of the paper titled MILD: Multi-Intent Learning and Disambiguation for Proactive Failure Prediction in Intent-based Networking, by Md. Kamrul Hossain and 1 other authors View PDF HTML (experimental) 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. Comments: Subjects: Networking and Internet Architecture (cs.NI); Machine Learning (cs.LG) Cite as: arXiv:2602.14283 [cs.NI] (or arXiv:2602.14283v1 [cs.NI] for this version) https://doi.org/10.4855...
