Pages

8 May 2014

THE ‘GOOGLIZATION’ OF INTELLIGENCE AND COUNTERTERRORISM ANALYSIS

May 5, 2014


The ‘Googlization’ of Intelligence and Counterterrorism Analysis

· By Dan Verton · May 02, 2014 

The ease of use and powerful search capabilities of Google have inspired a quiet revolution in the way some military intelligence organizations are using big data. But Google, the company, has nothing to do with it.

The company behind the big data search revolution in the military intelligence community is Melbourne, Fla.-based Modus Operandi Inc. Nearly a dozen military agencies, including the intelligence components of the services, have worked with Modus to conduct tests of the company’s Wave Exploitation Framework platform — a technology that leverages graph computing, natural language processing and big data analytics tools to help analysts map relationships between people, things, places and events.

Screen Shot 2014-05-01 at 2.59.39 PMThe Marine Corps and Army have recently conducted field tests using the Wave system, along with a new graphical user interface the company developed called BLADE, which is modeled after Wikipedia.

“It looks and acts like Wikipedia,” said Dr. Eric Little, vice president and chief scientist at Modus. “And you can search in a search bar just like Google.”

The familiar interface enables younger, less-experienced analysts to build profiles of individuals “just by clicking, dragging and dropping” information from multiple sources, he said. And it supports multimedia, including videos, photos and third-party apps.

Modus is marketing the technology to military, intelligence and counterterrorism organizations that are finding it difficult to manually discover relationships between terrorists and their supporters within the mountains of raw field data collected by soldiers and analysts. But the company believes the technology has much broader potential throughout the national intelligence community.

Graph computing

Screen Shot 2014-05-01 at 2.57.55 PMThe real power of Wave and BLADE comes from the ability to ingest massive amounts of data from existing stores and create an ontology that can be crunched and analyzed using graph computing.

First, the data is ingested and converted into a graph, a highly relational network model in which anything that can be classified as a thing becomes a node in the network. Those nodes are then assigned attributes for classification and search purposes.

For example, Wave enables analysts to understand the types of attributes that belong to persons, Little said. “People have parents, certain biometric characteristics, such as height, weight, hair color, eye color and skin tone,” he said. “They may have ancillary types of properties, like…glasses or a tattoo. You can describe certain features about where they live, where they work. You can also include things that people own, such as a cell phone.”

And that’s where things get interesting.

“Cell phones are the kinds of things that have phone numbers attached to them,” Little said. “Cell phones have call plans, cell phones have data plans, cell phones have carriers [and] cell phones have manufacturers. So you can see that from that person, you can go out to their cell phone; from the cell phone, you can go out to their data plan; from the data plan, you can hit their Facebook page; and from their Facebook page, now you’re connected to an organizational network of their friends.”

The graphs get very big, very fast. But when it comes to the ability to find things, Modus is actually using some of the same techniques and technologies Google helped develop. It leverages Hadoop and MapReduce to handle structured and unstructured data — the types of data that don’t fit neatly into database tables — and crunch it down into graphs.

Easy and smart

Modus has focused its efforts on making the tools easy to use, regardless of how much data they are taking in. In addition to ease-of-use, the company is now adding capabilities to make the technology smarter.

“One of the things we’re working on now is adding automated reasoning,” Little said. “What reasoning allows you to do is to build a logic over and against your graph, and that logic allows you to define certain kinds of rules, patterns and associations. And it will infer new connections and relationships” automatically.

“It will actually create data based on the knowledge that’s in your head,” he said.

To help prevent analysts from getting lost in a maze of far-flung connections and relationships, Modus is developing advanced search capabilities it calls “directed graph search, or faceted navigation.”

“In other words, every time you make a move and you’re traversing the graph, it tracks what you’re doing,” Little said. “It’s actually making a graph about the search of the graph.”

The point is to understand how you determined linkages in large, complex networks and to be able to share that workflow with other analysts, he said.

“You’re not only tracking the raw data in the graph,” Little said. “You’re tracking the providence and the pedigree of the workflows around that data.”

No comments:

Post a Comment