CAOE Quarterly Newsletter - October 2019

October 1, 2019

Are AI technologies biased?
New CAOE research looks to objectively evaluate AI technologies

Artificial intelligence is being embraced in both the private sector and across the homeland security enterprise. The use-cases for AI technology are wide-ranging, from helping users navigate immigration systems to predicting and preempting threat, to reinforcing critical infrastructure resiliency against possible attacks. CAOE researcher Mickey Mancenido, assistant professor, School of Mathematical and Natural Sciences at Arizona State University, is developing a testing methodology for evaluating the algorithm fairness of smart biometrics. In a recent TEDx talk, Mancenido discusses how biases can impact AI technologies and how this research can help DHS eliminate the technology’s potential biases.  
Click here for video.


Researchers at CAOE use narrative analysis to gain insights on international relations

Humans make sense of the world around them through the construction of stories. While different cultures offer varied perspectives, analyzing the stories told in foreign media can help researchers gain an understanding of diverse social and political environments, as well as assess how differing viewpoints impact relations with the U.S. The MESA group research team led by researchers at Oklahoma State University and Monmouth College have been using narrative analysis to evaluate conflict potentials on the Korean Peninsula, discover the motivations and perceptions of migration to the U.S. from Central American through media narratives and examine how global competition is discussed among nations taking adversarial positions to the U.S. The value of this research gives the Homeland Security Enterprise (HSE) contextual background and insights that allow for:

  • In-depth perspectives on relationships and interdependencies within communities, often those in conflict with the U.S.
  • Greater understanding of how messages move and change as they travel across communities. 
  •  Finding similarities in how problems are discussed across nations and finding common solutions to solving problems. 
  • Provide a structured look at how an actor understands itself and its role in the global community. 

In this video, Skye Cooley and Ethan Sample discuss a current project on the Northern Triangle and Mexican Media Perspectives on the Migration Crisis. 


CAOE adds new projects for year 3

The CAOE is excited to introduce several new projects, researchers and institutional partners to the center’s project portfolio for year 3.   

Deferring decision: Effects on human-AI performance

Artificial intelligence technologies, such as machine learning models used to match a person's biometrics with their identity, are increasingly being deployed for use in national security, defense and criminal justice. In these high-stakes applications, it is critical that human workers remain in the loop and that the joint human-technology system makes accurate, fair and responsible decisions. Therefore, it is imperative to understand how workers use this technology and the impact of these technologies on human performance. This research will evaluate the performance of human-machine systems in real-world use-cases and estimate the potential performance costs of implementing deferential AI with human decision-makers.

PI:  Mickey Mancenido, ASU 

Scalable near-real-time identification and characterization of malware behaviors using darknet data

Using data from previous domestic terrorist threats, researchers have created a framework to identify possible domestic threats and are developing training modules with DHS to teach first responders and local law enforcement how to better identify possible homegrown terrorist activities. 
PI: John Yen, Penn State University 

Validating adaptive behavior models of adversaries for risk assessment (VABMARA)

Constantly evolving adversary tactics, rapid technological advances and changes in the environment, especially concerning defenses, render a simple analysis of historical events inadequate for validating models that seek to capture fine-grained, tactical adversary behavior. This project will build a framework for validating models of adaptive adversary behavior through course-of-action red teaming (COART) by completing a comparative experimental analysis of game theory and different types of proxy human decision-makers. Partnering with Transportation Security Administration, this research will focus on the assessment of risk for adversary attacks against aviation targets.
PI: Gary Ackerman; University at Buffalo (The State University of New York)

“Moscow’s ‘trojan horses?’ elite risk-taking, disruption, and Russian strategic interests in German-speaking and east central Europe (GS-ECE)”

By examining the potentially pro-Moscow "Trojan Horse" effect of selected political parties, non-party grassroots organizations, business corporations and certain prominent former political actors who are still publicly active in German-speaking and East Central Europe (GS-ECE), this project assesses the possible effects of such “Trojan Horses” between 2020 and 2024 in light of the putative end of Russian President Vladimir Putin’s tenure in office. 
PI: David Dorondo, Western Carolina University 

A novel data-driven model for quantitatively evaluating the effectiveness of border security investment
While the U.S. northern border remains an area of limited threat in comparison to the U.S. southern border, the northern border presents unique challenges. The most common threat to U.S. public safety along the northern border is the bidirectional flow of illicit drugs. Other threats include economic migrations, asylum seekers and human trafficking. This project looks to identify, quantify and interpret the relationship among threat streams and the associated consequences of northern border risks. To develop a decision support tool to help HSE understand the potential costs and benefits resulting from operational changes such as changing border security countermeasure investments or increasing patrol agent staffing.
PI:  Jun Zhuang, University at Buffalo (The State University of New York)

Examining ideologically-motivated cyberattacks to better secure cyber critical infrastructure from compromise
 Most current research on ideologically-based attacks has focused on physical crimes or financial offenses enabled by off-line relationships. There is very little research on ideologically-motivated cyberattacks. This research identifies methods to reduce the risks of future attacks through better detection and defense techniques that can be implemented to improve the state of cyber-readiness and resiliency to cyberattacks.
PI: Thomas Holt, Michigan State University

A Flexible Framework to Achieve Strategic Objectives in US-China Policy

This project will assess current U.S.-China policy and identify strategic opportunities for innovation and engagement and provide a strategic framework to engage, challenge and integrate current U.S-China policy. It is anticipated that internal demand in China will force social change, ultimately resulting in domestic demand for political change. This assessment and projection will involve short-term (engagement), medium-term (challenging) and long-term (integrating) horizons for policy analysis for managing U.S.-China relations.   
PI: Maorong Jiang, Crieghton University

Chinese Cooperation and Competition with the United States: Cognitive and Motivational Analysis

This project will assess the integrative complexity and motivational hierarch of the governmental and (ostensibly) private-enterprise leaders involved in global commercial and political strategies of the People’s Republic of China. Analyzing open-source texts, this project will identify baseline and changing patterns of cognitive processes related to planning, decision-making, goal-setting and relations with other individuals and entities, as well as strategies for coping with stressful conditions. These analyses can lead to a better understanding and possibly better forecasting of Chinese political, industrial, technical and economical actions.
PI: Peter Suedfeld, University of British Columbia


CAOE students shine at COE Summit 2019 

The annual Center of Exellence (COE) Summit was held at George Mason University July 31-Aug. 1 and featured opportunities for CAOE students to showcase their innovative research benefiting Homeland Security. CAOE had students participating in the Summit Grand Challenge and the student poster showcase. The Grand Challenge, a new event for the Summit, was a student-led, team-based competition to "identify an emerging threat to homeland security posted by Unmanned Aerial Systems (UAS) and develop a strategy to counter it.” Six teams competed in this challenge, including CAOE undergraduate student researcher Emiliano Ruiz, University of Texas, El Paso, who was part of the winning "blue" team with the proposal, “Automated Drone Integrated Systems (ADIIS).” 

CAOE student researchers who presented posters also exemplified the innovative work being accomplished through the DHS Centers of Excellence. Students from Arizona State University, North Carolina A&T and University of Central Florida presented CAOE posters. Dominqueca Edwards and Sierra Marshall from North Carolina A&T represented the CAOE well, placing 2nd overall in the poster competition.  
 

 

Winning Blue Team including CAOE student Emiliano Ruiz (3rd from right) accepting awards for COE Summit Grand Challenge. 

Photos:Undergraduate students  Dominqueca Edwards (North Carolina A&T), Sierra Marshall (North Carolina A&T), Dan McCarville, CAOE Associate Director of Education, undergraduate student Gabriela Coccora (University of Central Florida) and recent graduate Brianna Chavez (Arizona State University) 

Upcoming Events


Distinguished Speaker Series
Identifying Key Objectives and Creative Alternatives for DHS Policy Decisions 
October 31, 12:00 - 1:00 pm at ASU



CAOE Biennial Meeting November 14-15, 2019
CAOE researchers and leadership will be meeting with DHS stakeholders and leadership in Washington D.C. for a two-day internal review of our first two years of research and operations.