Adaptive Digital Twin Synchronization: A Mechanism for Where, When, and What to Measure

Network Digital Twins (NDTs) require timely and cost-effective synchronization with physical networks. Existing telemetry and monitoring tools provide the plumbing to collect rich measurements and show potential for synchronizing the twin with its physical component. However, they lack a principled, fine-grained policy to identify and transmit only the necessary information, in order to save communication costs. In this paper, we present the idea of a Neural Measurement Field (NMF), a learning-based adapter that unifies where the measurements are taken (space/location), when they are updated (time), and what information is synchronized (content) into a single continuous intensity function, trained online to maximize decision utility under the budget.