U.S. Customs and Border Protection is setting up a data science division to improve its ability to monitor the border, according to a recent HSToday interview with the chief of the Border Patrol’s strategic plan and analysis directorate.
So far, only a half-dozen CBP staffers and an undisclosed number of contractors are developing the division, which Chief Benjamine Huffman said will gradually help more accurately determine the traffic that flows through the U.S. border with Mexico.
“We’ve never really understood how big a problem we’re trying to solve,” Huffman told HSToday. “We call it the the denominator problem. How many people are coming across the border, how do we really know how we’re doing. … So, our data stats division is going to start doing the analytics to help us get to those pieces. And we’re going to do a lot of simulation modeling.”
Chief Rodney Washburn, director of the CBP Data Analysis Division, introduced the initiative to Huffman, and will oversee its implementation. No other details have yet been made public regarding the program, its scale or required resources.
“I envision the Border Patrol will be able to classify things much quicker and faster in the future, relying on technology as things go, identifying what’s coming way before it gets to the border, being able to be there to secure that if all those things start working properly,” Huffman said. “So, our data stats division is going to start doing the analytics to help us get to those pieces.”
The Cyber Strategy
In May, the U.S. Department of Homeland Security, under which CBP operates, released a government-wide cybersecurity strategy – outlining a goal of shoring up the cyber detection, making tech improvements and hiring and redirecting staff resources. The CBP strategy was unveiled in 2016, and broadly outlines the goals of the organization.
Former CBP deputy commissioner and former Border Patrol Chief David Aguilar said system integration is the real tech challenge.
“We have unattended ground sensors, we have integrated fixed towers, we have most video surveillance systems, we have aerial platforms… We have the predators, we have Huey helicopters, we have Blackhawks and it goes on down. Each one of those platforms basically still pretty much operates independent of each other,” he said. “So, what the Border Patrol is looking for right now is how do we take, first and foremost, each one of those capabilities, and to the degree possible, independently use machine learning for that remote video surveillance system to tell me, ‘I can’t see what’s moving up there. But something is moving. Is it a car? Is it a truck? Is it a person? Is it a cow? Is a coyote?’ And if it’s a person or a car, what type of car? If it’s a person, are they carrying a long rifle or are they carrying backpacks?”
“That’s machine learning,” Aguilar added. “That’s one, and then how do you take each one of those independent machine learning capabilities and tie them in to where algorithmic targeting can tie in each one of those brushstrokes to paint that picture of what’s really happening out there. You go from domain awareness to situational awareness. That’s what they need right now.”
The Midterm Elections
President Trump said Wednesday at the White House that he wants to work with Congress to find funding for a 2,000-mile border wall along the southern U.S.-Mexico border. While CBP is directed by the executive branch, it is funded by Congress, and the 2018 midterms may put future border funding in jeopardy. Over the summer, the House failed to pass the Border Security and Immigration Reform Act of 2018, which would have provided funds for the proposed border wall, the construction of new border ports of entry and other improvements. The upcoming 116th Congress, with the House under Democratic control, will decide on the DHS budget for the remainder of the fiscal year.
“The battle over border wall funding could have a significant impact on the other parts of the DHS budget including cybersecurity, for everything from critical infrastructure protection to training to cyber threat detection and mitigation,” said a post-election day analysis by Cornerstone Government Affairs.
The Border Patrol, in fiscal year 2018, turned away 204,288 people who were deemed inadmissible or left the country after a short period, and temporarily detained 361,993 immigrants not lawfully in the U.S. This resulted in 566,281 total enforcement actions against individuals, an increase of 40,000 over the previous year.
At this moment, a caravan of approximately 3,500 migrants is headed to the southern border through Mexico to claim asylum in the U.S., prompting Trump to mobilize thousands of active-duty military troops to assist CBP officers.
“The ability to very quickly cross the border and assimilate into the legal population makes it very difficult (for CBP officers),” Aguilar said. “[In El Paso], 250-300 people are being released every day. In the Rio Grande Valley — double that, triple that. So, all of these people, remember, are coming into a broken system.”