We are pleased to announce a significant presentation from ONDM 2024 in Madrid:
Hussein Fawaz, Farhad Arpanaei, Davide Andreoletti, Ihab Sbeity, José Alberto Hernández, David Larrabeiti, and Omran Ayoub
“Reducing Complexity and Enhancing Predictive Power of ML-based Lightpath QoT Estimation via SHAP-Assisted Feature Selection,” ONDM 2024.
🔗 Access the paper here: https://doi.org/10.23919/ONDM61578.2024.1570996346
Key Highlights:
- Introduces an Explainable AI methodology (SHAP) to identify and select only the most impactful features for ML-based QoT estimation.
- Achieves substantial reductions in computational cost, inference latency, and telemetry data, while enhancing predictive accuracy.
- Offers a practical and efficient strategy for deploying ML models in real-time optical network operations.