INTERVIEW: David del Val Latorre, chief of Telefonica’s R&D unit (pictured), warned disparate data sets supplied by different vendors provided a huge challenge for operators attempting to use artificial intelligence (AI) and machine learning to automate network processes.

Speaking to Mobile World Live, Val Latorre said Telefonica had been working on homogenising data in its networks for two years as alternative measures are provided by different vendors, which also vary by market.

“The data in our networks is very different,” he said. “Some vendors provide the logs from the router or switches. Depending on the vendor they are different so making it [data] homogeneous so that the algorithms can work on something to make a prediction is a huge challenge.”

However, the executive noted the great potential of machine learning and AI for network automation to reduce costs, increase efficiency and improve customer service across both remote and urban areas.

“In new 5G networks the level of automation you need is beyond what we have done so far. So we need a new type of automation and that’s where AI comes into play. With AI, we can detect anomalies in the network and predict when a failure is going to happen days or even weeks before it happens.”

To view the full interview, where Val Latorre also discusses the advantage of open access technologies to increase connectivity in rural areas of Latin America, click here.