How can machine learning help the Department of Homeland Security confront both climate change and biological threats? That is what the agency’s Center of Excellence is trying to figure out.
These issues, as well as domestic terrorism and cyber attacks, are preoccupying the minds of DHS’ science and technology leadership. The pandemic is a recent example of new applications for machine learning and data analytics in national security.
DHS Chief Medical Officer Dr. Pritesh Gandhi said he is not worried about the agency’s capacity or technical expertise to respond to bio threats that may arise from adversarial nations or other bad actors. The question, rather, is how does DHS operationalize its current capacity and what is the “niche” for COEs moving forward.
“There’s an opportunity to do risk modeling and build risk matrices across all of the sectors – as a way to help us think through how we respond to bio threats, and how we actualize the resources that we have,” Gandhi said during the virtual DHS Centers of Excellence Summit on Wednesday, hosted by George Mason University. “I think there’s a need for that, that kind of strategic-level applied research, particularly again, as it relates to risk modeling or building out a risk matrix.”
Gandhi, who has only been in his role for just over 100 days now, said applied research should be part of a national risk architecture that covers medical countermeasures (MCMs) to bio threats and what are the opportunity costs associated with stockpiling resources.
“Because this isn’t just about technical expertise for decision making on MCMs – it’s about being able to understand risk and opportunity costs with the decisions you need to make, and having a rubric or risk matrix to help guide that. One that is based on research, I think would be very relevant,” he said.
Another medical opportunity would be the domestic opioid crisis, which Gandhi said could benefit from geospatial mapping of local clinical data to spot high concentrations or preponderances of depression – a common precursor to narcotics abuse. This also connects to domestic terrorism intelligence.
“And of course, you do this in a way that is data safe. You rasterized this data, you point shift the data, and so you protect the integrity, and you protect the privacy. But all of a sudden, you’ve got an idea of where there is burden from an economic perspective, from a social and in physical mental health perspective, from a drug addiction perspective, then you’ve identified areas that you can make intervention,” he said. “Not surprisingly, when you look at, you’re starting to look at this burgeoning research on domestic violent extremism, these challenges can come when you are faced with immeasurable poverty.”
DHS is enlisting its COE to combat domestic terrorism soft targets, especially in crowded places. Retired Coast Guard Rear Adm. Christopher Tomney, director of DHS’ Office of Operations Coordination, said his staff’s special events program works with state and local partners give events federal coordinating teams as needed. A key area for the COE is use of unmanned aerial systems (UAS) to surveil large-scale public events.
“And there’s a lot of challenges with dealing with how do you counter a UAS, when you have a large crowd below it, so you don’t create a flying piece of hardware that could come crashing down on the crowd,” Tomney said. “So, again, developing that technology to safely take control, technology that can quickly help to identify the individuals monitoring those, those UAS systems, when you have a large crowd underneath it.”
In addition, monitoring social media activity is nothing new for DHS. However, the COE can help find more efficient ways to comb through the mountains of data generated by tweets, posts, images and videos, etc.
“It’s real quick to get an alert that someone is tweeting something suspicious, but then we have to fall back into the traditional, labor-intensive way of trying to run to ground, whether or not that’s a legitimate threat or not,” he said. “So we are at the point where we are data saturated. So again, helping us to develop technology in which to immediately pull our attention to truly what is most important out there in cyberspace, versus all the other chaff that may be out there.”