Military intelligence, AI style: MoD cosies up to Massive Analytic

The Ministry of Defence (MoD) has received a “milestone” report from Massive Analytic, a company developing an “artificial precognition and decision-making support [system] for persistent surveillance-based tactical support.”

Massive Analytic, as we reported last year, is an automated analytics startup – and one of 24 SMEs that were contracted by the MoD in January as part of the British military's new innovation challenge.

The award came as the MoD looks to SMEs, rather than the historical big players, when it comes to military procurement, to develop tools and techniques allowing for the automated processing of data collected by its surveillance systems.

While the firm is based in London, with its flagship offices in Covent Garden, Massive Analytic's developers actually work at its Innovation Warehouse in Farringdon, where they integrate the data acquired through the MoD's persistent surveillance systems into an analytics system touted as providing a view of the future in battlefield scenarios.

Jim Pennycook, head of operations at the MoD's Centre for Defence Enterprise (CDE), told The Register: “The amount of data both required and produced by defence systems and processes is rapidly increasing and becoming more difficult to manage.

“In a time when military manpower is limited,” added Pennycook, “[the]manual processing of data is too time consuming. The use of autonomous systems and processes to make sense of data to support decision making could increase efficiency and reduce the risk and cost of operations.”

As one may infer from Massive Analytic's elephantine logo, its predictive and textual analytics platform, Oscar, runs on Hadoop. For the MoD contract, it will process data from autonomous vehicles, sourcing and validating big data in difficult environments and providing a dashboard to visualise and manage the information being collected.

Talking to The Register, Massive Analytic's CEO George Frangou explained: “In a surveillance scenario – perhaps in peace time or perhaps in a theatre of war – you will have lots of sources of information coming in. There'll be stuff from social media, there'll be geographic information, you'll have information about where assets are and where threats are, and information about how threats are going to be manifesting in terms of timing and geography.

“That information will come in different forms; some of it will be images, some video, some data, some text,” said Frangou. “The question is how can you get the holistic view from these things to precognise possible outcomes.”

In their report delivered to the MoD last week, Massive Analytic had established what the user-level requirements for its PoC were, with a view to “developing high level precognitive command and control [before moving on to developing] the integration points.”

As Frangou said, the MoD has got “processes and infrastructure in place for persistent surveillance. Our objective is to take the Oscar platform and develop a specialist front end for it, for surveillance. and integrating it within that infrastructure. We're building precognitive capabilities into command and control.”

Within traditional command and control systems that involve the tracking of assets, “some of the biggest challenges you have is multiple platforms, and moving data across multiple platforms with different protocols and interfaces” said Frangou. “If you can imagine a command-and-control structure with troops on the ground, ships at sea, or planes in the air – trying to coordinate those different platforms is difficult, so we're bringing them together.

“All of this data is very, very large,” said Frangou, “and having Hadoop's parallel processing environment makes it all easier. If you then have the interfaces to bring it into Hadoop then that allows you to avoid data warehousing structures.”

Fluidity is key, suggested the CEO, who extolled the value of Spark when streaming and analysing petabytes of unstructured data. Spark and Hadoop allowed the company to bring all of this data together and analyse it at scale, he added.

As Pennycook told us, the CDE was seeking out “anyone with an innovative idea that has a potential defence or national security application. In focusing on innovative technologies our investments in science and technology research are high risk with high potential benefit.” ®