Machine learning (ML) can be applied at the ‘edge’ where data processing functions are tightly targeted and power consumption must be low.

When machine learning algorithms are fed a steady stream of real-time data, they can quickly detect bottlenecks, anomalies and other operational inefficiencies.