Accomplishments

Founded in 2013, the Laboratory for Analytic Sciences launched with the goal of joining together professionals from academia, industry, and the intelligence community to provide solutions to real-world problems. From keeping attendees safe at the 2016 Rio Olympics to improving how humans and machines work together, LAS works to incubate the best ideas and bridge the gap between the art and science of analysis. The highlights on this page are a small selection of the accomplishments LAS has achieved since its founding.

Combating Disinformation and Malign Influence with Social Sifter

Just four years ago, Russia had to work hard to influence the U.S. presidential election, “pushing it along,” as national security expert and LAS performer Clint Watts said in mid-May on “Meet the Press.” But today, what he calls the “kindling” for wayward international influence—particularly from Russia, China, and Iran—is home-grown, with fringe groups taking up agendas, sometimes unknowingly, and spreading disinformation through social media to deepen America’s divide and erode trust. The uncontained spread is the “gasoline” poured on the kindling, causing what Watts calls the “Bonfire of Disinformation.”

Yet, wading through the sheer volume of material on the Internet is an arduous and slow process for intelligence analysts—one the Laboratory for Analytic Sciences is trying to make easier. The “Social Sifter” interface narrows the field, finding key users and accounts related to coordinated efforts to spread misinformation. Operating as a search engine, it applies an analyst’s criteria and filters for a variety of factors like origin, text, and language. LAS partners with experts like Watts and others who work in language, marketing, statistics, psychology, and a variety of other fields.

“Visualizing” Research in Real-Time with Navy Reservists

In the summer of 2019, the Laboratory for Analytic Sciences sponsored R. Jordan Crouser, an Assistant Professor at Smith College in the Department of Computer Science, and Alvitta Ottley, Assistant Professor at Washington University in St. Louis in the Department of Computer Science, to test a visualization theory in a real-life simulation with Navy Reservists: Did individual differences in personalities affect how analysts found information?

The reservists spent two days working on a simulated kidnapping case, with Crouser and his team of undergraduates from Smith, a historically women’s college, monitoring how each reservist approached the task. As reservists explored data, their interactions were automatically logged and categorized.

Before the simulation, reservists were given tests to determine if they had an internal locus of control (they had control of events) or an external locus (events happened without their influence). Ottley and Crouser found that those with an internal locus had more “unique clicks,” meaning they covered a broader field, but the question remains as to whether “more” is “better.” The research could enable the development of personalized and adaptive tools to better support intelligence analysis.

“Investigating the Role of Locus of Control in Moderating Complex Analytic Workflows” is funded by the LAS, an ongoing collaboration between the Human Computation and Visualization Laboratory at Smith College in Northampton, Massachusetts, and the Visual Data Analysis Group at Washington University in St. Louis, Missouri. Watch this project video to learn more.

LAS Presents Real Problems and Finds Real Solutions.

The LAS launched in 2013 with a primary mission: collaboration.

But getting professionals from business, government, and academia to share ideas wasn’t necessarily easy. How could the LAS improve the likelihood that collaboration would be productive?

Bev Tyler, Sharon Joines, and Jessica Jameson, NC State faculty from the Poole College of Management, the Design School, and the Department of Communication, and Kathleen Vogel, from the University of Maryland School of Public Policy, are part of a multidisciplinary team that examined LAS research models, publishing their findings in a series of papers that culminated in a book released in May 2020.

“In the early days of LAS, all participants were trying to figure out their role and how they could make the best contributions to the LAS mission,” writes Jameson and Joines in Chapter 5 of Facilitating Interdisciplinary Collaboration Among the Intelligence Community, Academy, and Industry. “A programmatic structure that required each contractor (faculty or industry partner) to have an independent deliverable, compounded by the zero history among most LAS members, reinforced a tendency toward working independently rather than in direct collaboration.”

The team cited several case studies from LAS that resulted in innovative and long-term outcomes; in fact, intervention from the team was no longer necessary after 2015. Some key takeaways needed to support strong collaboration:

A group leader who understands the needs of various partners and can serve as a bridge across the cultures.
Taking time to discuss and align language/terminology and identify key events and issues.
Sharing case studies, examples, and stories to help create shared meaning and experiences.

These objectives were achieved over time through many events—from “kick-offs” to introduce goals to “sprints,” which challenged groups to two-hour work periods using design theory methods and techniques to explore specific analytic processes. These kinds of events led to more productive teams that were able to work together to achieve products and outcomes of interest to the intelligence community.

In May 2020, the team published its findings in the book “Facilitating Interdisciplinary Collaboration among the Intelligence Community, Academy, and Industry.”

Development of GazeGIS prototype

Laura Tateosian from the NC State Center for Geospatial Analytics has developed a gaze-based reading and dynamic graphic information system (GazeGIS). Location is an important component of a narrative. Mapped place names provide vital geographical, economic, historical, political, and cultural context for the text. Online sources such as news articles, travel logs, and blogs frequently refer to geographic locations, but often these are not mapped. When a map is provided, the reader is still responsible for matching references in the text with map positions. As they read a place name within the text, readers must locate its map position, then find their place again in the text to resume reading, and repeat this for each toponym. GazeGIS uses eye tracking and geoparsing to enable a more cohesive reading experience by dynamically mapping locations just as they are encountered within text.