Government Cybersecurity Research Explores Technological, Human Capabilities
NSA labs probe chip-based devices and enhance cognitive learning to secure the nation.
Challenges ranging from teaching people new ways of learning languages to providing security for homemade computer chips head the priority list for researchers at the National Security Agency. The exponential expansion of technology capabilities is perhaps matched by the growth of potential conflict areas, and both are increasing the issues faced by the agency’s research community.
Traditional skills such as translating communications intercepts now must take into account that any one of thousands of languages spoken on Earth could be vital if a new trouble spot flares up. The ubiquity of networked devices, especially in the context of the emerging Internet of Things, provides its own unique cybersecurity challenges. And, the near future may see individuals making chips at home for their own customized communications devices, which also would need to be secured.
These are some of the tasks facing Dr. Deborah A. Frincke, director of research at the National Security Agency/Central Security Service (NSA/CSS). Frincke points out that hers is the only group within the intelligence community that has a large body of long-, medium- and short-term research. In addition to conducting contract research with academia and industry, it also has a sizeable investment in long-term staff. “We have a very large body of professional researchers who have spent their entire careers here and also those we hire more later [in their] careers … and that is unique,” she offers. “We don’t see that elsewhere—that investment in a long-term body of internal researchers.”
This institutional knowledge provides a significant advantage, she continues. With the research directorate inside the agency, it “sits right at the table’ with the senior leaders of the agency. Leadership hears about technological advances at the same time it is learning about worldwide issues, she points out, which enables real-time coordination of research with mission needs.
Other research agencies such as the Defense Advanced Research Projects Agency (DARPA) and the Intelligence Advanced Research Projects Agency (IARPA) cycle out researchers to ensure a steady flow of fresh blood. Not only does Frincke’s directorate have both fresh blood and long-term intellectual capital, its work with IARPA and DARPA is complementary, she allows.
“IARPA is a great example of a partnership where I can leverage their ability of really rapid rotation and their ability to be out and about in the community in a much broader way,” Frincke states. “There will be projects we do with IARPA, for example, where we’ll jointly decide that there’s a particular research area that’s best conducted on the outside but informed by internal knowledge.”
For example, the NSA’s long-term agenda in a field such as analysis might help drive IARPA decisions about where to invest and where to dedicate its unique resources, she continues. Both agencies then benefit from the results.
At the top of the NSA/CSS critical research list is language analytics. Frincke points out that the world today is rife with potential hot spots, and about 7,000 languages are spoken around the globe. A hot spot might arise amid any one of those languages.
But this effort involves more than just having a lot of people who speak foreign languages, which Frincke emphasizes the NSA already does as well as any organization. The agency teaches its team of linguists about 100 languages as part of its education and training, but that means more than 6,000 other languages remain out of the realm of the agency’s expertise.
With the NSA gathering information in an increasing number of languages, the agency is looking to further its expertise in human language translation. This research aims at helping a machine better assist the understanding of a conversation, she allows, and it departs from conventional language understanding programs.
“This is different from the [Apple] Siri approach where you perhaps have a language translation that’s wanting to be understood,” she explains. “It’s more like the cocktail party issue where you walk into a room and there’s all kinds of languages taught. How do you understand the particular one you need?
“That’s a much harder problem than is faced by the outside, where it’s generally one language at a time with a willing listener and a willing speaker,” she elaborates.
This effort relates to a second research priority—how to teach people to pick up a language more quickly, which ties into learning approaches, Frincke says. Many advances have emerged both from NSA research and from the broader activities of the NSA/CSS. One benefit that has emerged from a university research program operated through the University of Maryland reaches across several universities and other communities—the Center for Advanced Study of Language.
This benefit involves teaching people how to transition from one language to another. For example, a person may know a language similar to that of a region that has just emerged as a hot spot. That person theoretically could learn the new language quicker because the individual already has a degree of familiarity.
The language effort already has uncovered some interesting aspects of learning, Frincke relates. Human learning can be improved by enhancing certain kinds of cognition, she reports. Enhance those aspects of human cognition, and the person learns a language better.
“Early indications are, for example, that if we invest time in improving people’s short- and long-term memory—and we spend five or six weeks on that project alone to enhance their overall skills at learning and memory—they’ll learn a language faster than if we spent that time teaching the vocabulary,” Frincke declares. “So, priming the brain to understand language is a piece of what we need to get right.”
This approach involves the issue of “how do we watch the brain while it’s in action as it’s learning a language,” Frincke points out. “That is fascinating.”
A third research priority is cyber. Frincke admits that many different aspects of cybersecurity are important for the agency. Research efforts must take into account the intelligence community perspective, the information assurance perspective and work with the U.S. Cyber Command.
On the information assurance side, research is focusing on several key areas. One of them entails the basic concept of the science of security, she allows. Much of this work traditionally has been done ad hoc—problems are discovered, patches are applied and experts then hope no one exploits vulnerabilities until the next threat manifests itself.
To go beyond this approach, research aims at the underlying principles that are important in cybersecurity. Frincke offers that some new approaches may help inform better decisions long-term. Similarly, some radically different approaches to cyberdefense could be on the horizon. “This is an area where I can take advantage of the long-term ability of NSA research to focus and put some of our researchers behind that area,” she states.
The NSA is expanding its science of security paper competition, in which it reaches out across the country to identify best practices that it sees in academia and industry. The agency also is funding “lablets” that would both advance the science of cyber and bring in more people who would address problems with a clean slate. This approach aims to use a zero-based build to design cybersecurity systems from the bottom up without the existing hurdles wrought by existing technology and legacy systems.
The agency’s Emerson Building houses what Frincke describes as the intelligence community’s first classified and unclassified wireless capability, including an Internet café. The NSA may be able to learn new means of wireless defense by using this facility as a testbed. Resilience is a key approach to cybersecurity. Frincke allows that the agency is “exploring with great enthusiasm” how defenses might be designed into a system so it could be self-defending. This work includes examining the pros and cons of this approach as well as how effectively it might be used more broadly, she says.
Addressing information assurance from a science-based approach builds a foundation on which other kinds of communications media can be examined, Frincke offers. With all types of new media and related technologies emerging, the NSA/CSS must be able to defend any kind of communications device. Experts must determine which pieces of a device must be protected while preserving the privacy of the user and without affecting the integrity or function of the device.
“We can emphasize in our long-term research of the principles, and then we need to have an expertise that’s able to quickly understand new technologies,” she says.
She cites 3-D printers as an example of an exploding technology that will tax the expertise of technology organizations. Before long, people will be able to produce chips with their home printers. Users will be able to individually tailor their own information technology devices, and those devices must be protected. That security must be principle-based.
Another onrushing advance that challenges NSA/CSS researchers is the Internet of Things. “When you think through what might be a blank slate protection of a home,” that brings up several issues, she points out. “How might we better defend privacy at the same time we’re providing security in a home system?” Frincke adds that the NSA/CSS has some unclassified research in that area.
Some of the security research into resiliency could be applied to the Internet of Things, she points out. Developing a self-aware network that can defend itself requires the ability to share information for mutual protection from one aspect to another. The challenge is to make a large number of devices self-defending and coordinating when the original structure of the device is not known.
“So, [it’s] a bit of a black box approach,” Frincke says. “How do you help that black box protect itself in conjunction with others? That’s fascinating, and a lot of breakthroughs—a lot of basic science and a lot of near-term research—are being done in that area. I think that is going to come up with some really nice transformational results for us.”
Coordinating NSA/CSS research with industry is difficult, Frincke allows, especially because “industry doesn’t really coordinate with itself.” Multiple approaches underway offer the potential for successful efforts with industry. One she cites is outreach and public conversation. “As much as possible, we are taking our research agenda externally and discussing it in an open and public environment—in as much detail as we’re allowed, given the circumstances,” she says, alluding to the classified nature of much of the agency’s research.
Another approach is to seed the security community with information about the agency’s security challenges. This includes presenting a high-level view of some of the research underway along with a question-and-answer session on future areas of complementary investment. The agency also teams internally to work with groups such as In-Q-Tel that are involved with industry.
Outreach to academia also fuels NSA/CSS research requirements. Frincke relates that her directorate has opened a new laboratory at North Carolina State University known as the Laboratory for Analytic Sciences, or LAS. Researchers assigned to the laboratory by the NSA/CSS are working in a secure facility with academic and local industry partners.
The effort at LAS also is examining how the NSA/CSS can team better, she adds. Part of this analysis involves determining which projects are of interest to the NSA/CSS and the cadre of researchers at the LAS. Work might be performed within the LAS, or it might be outsourced to another facility that would be provided with key data from the LAS. The NSA/CSS and IARPA have been planning this effort together, Frincke relates.
Another project from the NSA/CSS and IARPA involves how serious gaming can better inform teaching and learning methods. Frincke offers that IARPA launched this research project, but the NSA is leveraging its findings for its own education and training systems.
Both IARPA and the NSA are looking at exchanges of personnel, Frincke states. She says this might be a detailee who would be in a research billet but working for IARPA Director Peter Highnam. This would allow that individual to be engaged more directly in some of the more sensitive research that the NSA/CSS needs to keep at arm’s length but always have on tap, she says.
Frincke allows that she is assessing how her directorate is using its resources, particularly in light of the current budgetary climate. She relates that she spent much of her first months identifying activities and assessing how funds were being spent. She now is in the position of ensuring all funds are spent properly and every billet is deployed correctly.
“We’re conducting the first ever scrub from top to bottom of the research directorate, looking at every project that we have in the context of the big picture of ‘what does the nation, what does the NSA, need from the directorate of research?’” she states. “We’ll be spending a lot of time thoughtfully examining each project, seeing what needs to be extended, thinking about which ones I might need to pull off the table, making some hard choices—and I’ll identify the gaps.
“So, once that’s done, I’ll know if there’s a need for additional resources,” she continues. “But my goal now is to live within my means.”