Vanguard Magazine

20 JUNE/JULY 2017 www.vanguardcanada.com JoInt TArGETING J ing, identifying, tracking, and engaging targets due to the in- creased data flow, the sophistication of adversaries in concealing the data, and the speed of conflict. The expression, "time is of the essence" has never been more true. In today's operational theatres where our adversaries blend in with the local population, they hold an advantage by only being able to only briefly expose them- selves to large-scale military operations. The momentary need for exposure, often prompted by the need to communicate, occurs in an instant. When adversaries take advantage of the ease of light- ning speed of communications, they leave a trace. Bread crumbs can be followed, dots can be connected, and machine learning algorithms is the way to do it. Recent operations have demonstrated the pressures and benefits that data science can bring to the Canadian Armed Forces Joint Targeting process. Whether analyzing maritime targets while in command of Operation Enduring Freedom's Combined Task Force 150 or providing targeting solutions for Operation IM- PACT and the global coalition against ISIL, we can begin to miti- gate human bias and our finite intellectual capacity by leveraging the exponential growth and availability of data. Denying terrorists the use of the sea to transport drugs or denying ISIL the revenue from oil production is complicated military business. Relation- ships, money, and communications methods all play a role in these kinds of networks, and machine learning on big data can make the connections. So far we have discussed Joint Targeting for the CAF as something new and a clean slate to take advantage of, but that is not entirely ac- curate. The CAF constantly seeks to improve itself and so for several years has been implementing concepts and ideas that turn out to be strong starting points for a Joint Targeting enterprise. In 2009, the Canadian Forces Integrated Command Centre (CFICC) was established for the operational commands that were transitioned into the Canadian Joint Operations Command (CJOC). The Integrated Command Centre provides an opera- video, text and anything posted in social media, for example. Big data is collected and saved in the cloud by personal mobile de- vices, satellites and remote sensors, software logs, cameras, micro- phones, radio-frequency identification (RFID) readers and wire- less sensor networks, to name some. Whether you are going to the grocery store, streaming a movie or out for a bike ride, virtu- ally every activity and query creates recordable data. Those kinds of details saved in massive data lakes mean that there are endless possibilities when posing questions and seeking answers. As the Economist noted recently, Alphabet (Google), Amazon, Apple, Facebook and Microsoft all deal in data. In this digital age, data is the number one commodity, not oil. In terms of Joint Targeting, what do we want out of these data lakes? What we want is an ability to trace kernels of information and reveal hidden patterns, particularly those relating to human behaviour and interactions. The business world does this with ma- chine learning algorithms, and the same science and technology could be adopted to facilitate a targeting process. Artificial intelligence has brought us computers that can beat us at a game of chess, process human speech and read handwriting. But there has been even more impressive progress in the past few years in the artificial intelligence subfield that we have already be- gun discussing – machine learning. Machine learning algorithms help computers map rules and find loose connections by them- selves without having to be expressly programmed by a human and are the key to uncovering the buried information in big data that leads to useful intelligence. With machine learning, computers can predict when a situation is coming to head in a foreign land, predict weather systems on a continental scale, and scour all available medical literature to uncover cancer treatments for more obscure forms of that disease. The real beauty of machine learning is that the more data accessed by a machine learning system, the more it learns to do its job bet- ter. Freed from the finite capacity of the human mind, machine learning is able to discover the patterns buried in the big data that can be useful to a Joint Targeting Cycle. The Canadian Armed Forces' and NATO definition of targeting is: The process of selecting and prioritizing targets and matching the appropriate response to them, taking into account operational requirements and capabilities. Simply put, a target can be a per- son, place or thing. The targeting cycle is often presented as it- erative – analyzing and prioritizing munitions and non-munitions capabilities against targets – and consists of six major steps: (1) end state and commander's objectives, (2) target development and prioritizing, (3) capabilities analysis, (4) commander's deci- sion and force assignment, (5) mission planning and force execu- tion, and (6) assessment. Each of these steps is driven by subject matter experts, intelligence analysts and operational planners. As a team, they execute the targeting cycle as efficiently and precisely as possible. Today's warfare brings with it significant challenges in detect- The real beauty of machine learning is that the more data accessed by a machine learning system, the more it learns to do its job better.

• Cover