Policy-driven Conformal Prediction for Trustworthy QoT Estimation

2026-06-10

AI Summary

This work addresses the limited reliability of Quality of Transmission (QoT) estimation under distribution shifts and its misalignment with operational decision-making by proposing the Conformal QoT framework. For the first time, it integrates a policy-driven mechanism into conformal prediction and combines it with domain adaptation techniques to simultaneously ensure statistical validity and align predictions with real-world operational requirements. Evaluated on a public dataset, the proposed method improves prediction accuracy from 92% to 99.6%, substantially enhancing model robustness and practical utility in the presence of distributional shifts.

Abstract

We propose Conformal QoT, a policy-driven framework that combines statistically guaranteed QoT estimation with operational decision policies, enabling reliable lightpath-feasibility predictions under domain shift and improving accuracy from 92\% to 99.6\% on open datasets.

Problem

• Conformal Prediction
• QoT Estimation
• Domain Shift
• Lightpath Feasibility
• Trustworthy Prediction

Innovation

• Conformal Prediction
• Policy-driven
• QoT Estimation
• Domain Shift
• Lightpath Feasibility