Generative AI for Network Operations (OFC 2025 Workshop)

Problem Statement: Forecasting traffic demand is heavily reliant on traffic actuals collected from network. Collection gaps due to outages, misconfigurations, timeouts, etc. result in incomplete traffic actuals affecting forecast quality. Solution: Impute missing and anomalous traffic data to ensure continuity and the quality of our traffic demand forecasts using Generative AI models Approach: Implement a GAN model to improve data quality by correcting anomalies, filling missing values, and generating realistic synthetic data to augment existing datasets