Securing borders and supply chains and improving efficiency: Q&A with Ross Maciejewski

Following the terrorist attacks of 9/11, the U.S. government invested heavily in security and public safety. The increased focus on security created a huge influx of information, and the government began looking for ways for decision-makers and workers to quickly interpret and analyze that data.

Visual analytics and data science emerged as the answer. Just as these fields became a national priority, Ross Maciejewski was studying them in his graduate work at Purdue University. He soon earned a scholarship with the Department of Homeland Security and continued to conduct applied research to employ visual analytics to help law enforcement.

Now, he leads the Center for Accelerating Operational Efficiency, or CAOE at Arizona State University’s Global Security Initiative. Founded in 2017, as one of nine Department of Homeland Security Centers of Excellence, CAOE develops and applies advanced analytical tools and technologies to enhance planning, information sharing and real-time decision-making in homeland security operations.

“Our center is unique in the Center of Excellence network in that we're the broad crosscutting center,” says Maciejewski, also a professor and director of the School of Computing and Augmented Intelligence. “We have four pillars that include data analytics, operations research, economic science and homeland security risk science. Our underlying goal is to improve operations using different techniques from all these different fields.”

In recognition of GSI’s 10-year anniversary, Maciejewski looks back on CAOE’s successes and growth, and their continued efforts to equip DHS with better tools to ensure public safety. 

Question: Why is CAOE’s research important to national security and to society overall?

Answer: One of our greatest successes has been our Transportation Security Administration Plan of Day Scheduler, which uses real-time data collection and analytics to predict passenger arrivals. This allows the TSA to better allocate staff and resources where needed. One reason this is so important is the cost savings. If we can find places where we can reduce costs for the federal government to keep security levels at or better than they currently are, this is a way we can really have an impact.

Within the Center for Accelerated Operational Efficiency, we're always looking for these places where we can save resources, whether it's through thinking about how to improve throughput or thinking about how we can improve analytic tools so that analysts can do their jobs better, faster.

And being able to help the federal government in these operations and keeping national security at the forefront of things is an exciting opportunity to live up to that ASU mission of use-inspired research and real-world impact.

Q: And in terms of benefit to society, we all like shorter lines at the airport, right?

A: Shorter lines at the airport — but also, we have projects on reducing elicit goods, such as trafficked opioids, being smuggled into the United States. We've had projects with the Procurement Innovation Lab to examine how can we make contracting go smoother. If you can save some money during the contracting time, that can add up to a whole lot of money. We have people that are working on human-AI teaming, ensuring workers are feeling satisfied in their jobs while teaming with AI components.

We're working on a lot of ways to improve our national security footprint so there’s less hands-on intrusion, but security standards are still being met.

Q: How are students involved in CAOE research?

A: For most of the general public, when they think of DHS, they probably think of the most visible, public-facing roles such as Transportation Security Officers at the airport or Border Patrol agents. But there are so many more jobs that require engineering and science backgrounds in these agencies and we get STEM students acquainted with them through experiential learning options.

For instance, the DASSH challenge — Designing Actionable Solutions for a Secure Homeland — is a design challenge we run with another Center of Excellence called SENTRY. We source challenge statements from the DHS and then we bring together students to iterate on ideas on how to solve these problems.

The students get to ask domain experts and scientists questions, form teams and work on these projects, and then present to a panel of experts and get judged and win prizes. I think we had 112 students involved last year.

We also spin that off into a summer experience for quantitative analytics. It’s a four-week summer program for undergraduates that targets students from minority serving institutions. Students work on different data projects, learning data collection optimization, data mining and machine learning visualization, and combine all these together into a four week long project with us.

This is in addition to our broad portfolio of research projects — anywhere from 10 to 20 during a given year — that all involve graduate students, and many of them also involve undergraduate students. 

Q: What are some emerging or pressing challenges and how is CAOE addressing them?

A: AI, of course, is at the forefront of every news article lately, and one of our current challenges is how can we take some of these generative AI techniques and use them to improve efficiencies. They're pretty good at document summarization, they're pretty good at helping with different workflow patterns, but they still have issues like hallucinations. So, you can imagine an intelligence analyst has a plethora of intelligence reports and they want a quick summary to look for certain information. They might put those into a large language model, but you want to ensure the report is accurate and not having a hallucination.

That may be a bridge too far right now, but there are still scenarios in which these tools are useful. For instance, imagine a county emergency manager needs to develop an emergency response plan. It’s a very underfunded county, very small rural population, and the neighboring counties probably have similar issues. If they've created a plan, an LLM could help walk the manager through creating their own without having to create everything from scratch.

Another emerging issue is supply chain disruption. COVID-19, the war in Ukraine, the cargo ship that got stuck in the canal, the bridge failure in Baltimore — all these things lead to supply chain disruption. We've had a broad interest in bringing together operations research, economic analysis and risk science to think about how we can make supply chains more resilient.

Q: COVID-19 presented some unprecedented challenges to the global supply chain, but disruptions such as war and shipping issues have always existed. Is there something different about the global logistical landscape that makes these age-old challenges more difficult for governments and nations to grapple with?

A: As time has passed and we've emerged into this global market, we're reliant on all these different resources coming and going from different places. Now a single point of failure can lead to large-scale disruptions. All these ports need to operate at 95-100% capacity, and anytime there's a disruption, then you wind up with this huge backlog.

So right now, we’re just trying to think about where these disruptions might come from, help develop smart planning for these sorts of things, and give people tools for this risk assessment to try to come up with backup plans. I think this is critical and interesting work. A lot of historic planning has involved a single point of failure, but now we have researchers that want to look at cascading events. If this happens, then what cascades out and how do we understand the second, third, fourth order effects on this disaster?

Q: Can you tell me a bit about the relationships and partnerships that propel CAOE’s mission and research?

A: The center itself is well designed to facilitate broad, cross-cutting interactions. We have long running projects, but as director, I also routinely meet with DHS leaders to learn about their problem spaces, which we turn into calls for proposals to tackle these issues.

In this way, we can not only be responsive to new needs, but we can pull in experts from across the country to expand our network. We’re not locked into partners in our existing system; we can get the best people in a given area across the country to come join us. 

We bolster this with seminars featuring great speakers. For example, artificial intelligence is in forefront of people's minds at the moment. We have a generative AI series related to the Homeland Security Enterprise. Our speaker in December was Ed Chi, the vice president of research for Google DeepMind. So we're getting tons of interesting people coming to talk about these topics and think about how we can all work together in a smart manner.

Q: Is there something about GSI’s culture or structure that helps facilitate these connections that have made CAOE such a success?

A: All security research interests at ASU roll up into GSI, so there’s a structure that can see the broad picture and make connections. For example, sometimes we’ll have visitors on campus from the federal government, and GSI can connect them with CAOE or any of our other centers if interests align.

But overall, I think the Global Security Initiative has done a really nice job of trying to galvanize this broad concept of security across ASU. It has been very valuable having that resource that knows how to interact with places like DARPA, the Department of Defense and DHS, and looking at the security landscape across ASU.