Skip to content

Big Data and Diverse Threats: The Essential Role of Domain Expertise

July 9, 2013

Dr. Andrea Little Limbago 

The recent presidential campaign placed data analytics in the limelight, highlighting its growing role across academia, industry, and politics. Whether it was the use of data scientists to mine vast amounts of information about voters, the pre-election vitriol for Nate Silver’s predictions for an easy Obama victory, or the post-election glorification of all-things data, quantitative data analytics is back en vogue after the scorching it received at the height of the recession. This trend has finally made its way into national security community. For instance, the recent Foreign Affairs May/June cover story, appropriately titled “The Rise of Big Data”, notes the transformative effect of the big data environment.

As is often the case, the pendulum swings sharply, with a growing movement toward big data analytics and away from domain expertise. Machines can simply replicate the knowledge inherent within a domain expert, it is argued, and provide more provocative assessments in a shorter period of time. In fact, a panel at the Strata 2012 Conference (a major conference in the IT world) debated the value of machine learning versus domain expertise. Not terribly surprising, an audience full of computer scientists and engineers sided with machine learning over domain expertise. Similarly, there are articles that point to the end of theory, as well as the all-predictive power of big data.

Unfortunately, these arguments mask the limitations of big data, which must be addressed given the high stakes involved in the national security community. Too often, many of the computational analytic solutions within the national security community lack domain expertise –in the choice of data, the theories underlying the models, and the results interpretation. This does not mean that big data analytics cannot be useful, but that they must be married with domain expertise. The national security community – given the diverse threats and high stakes – should in fact take the seemingly bold move and integrate domain experts with the big data solutions. Given the broad array of threats, as well as the inherent limitations in big data, domain expertise should be viewed as a necessary condition for any technological solutions aimed at augmenting the analytic tradecraft.

Shifting Environment

For the national security community, domain expertise broadly refers to the wide range of analysts – found in the Department of Defense, the Intelligence Community, think tanks, academia and industry – who possess subject matter expertise in areas ranging from security studies to regional studies to foreign policy. Unlike during the Cold War where there was a significant emphasis on Kremlinologists, today’s domain experts cross a broad range of social sciences and regional expertise. As Director of National Intelligence James Clapper notes in the 2013 Worldwide Threat Assessment to the US Intelligence Community, “Threats are more diverse, interconnected, and viral than at any time in history.” He reinforces the perspectives previously put forth in other strategic doctrine, such as the Quadrennial Defense Review and the Joint Operating Environment. Each document describes the dynamic and multifaceted environment, which contains a multi-polar distribution of power coupled with challenges of weapons of mass destruction, cyber, the persistence of violent extremist organizations (VEOs), as well as demographic shifts, natural resource tensions, and the effects of climate change. These challenges all converge within an era of globalization wherein technological advances are reshaping the status quo across the globe. The information technology revolution brings with it new challenges, empowering non-state actors to exploit vulnerabilities with very limited resource requirements, while also introducing the cyber sphere as a new domain for warfare. The diverse range of challenges in the global environment renders domain expertise even more pertinent, especially given the necessity to interweave and prioritize strategies and programs across the defense, development, and diplomatic spheres. As Dan McCauley recently noted in Small Wars Journal, “Given today’s dynamic and information-laden strategic environment, senior leaders cannot possibly possess the depth and breadth of information essential for informed decision making.”

Big Data and National Security

Not only is the national security community juggling a range of threats, it also is drowning in a sea of disparate data. From the White House’s overarching Big Data Initiative to the DoD’s Data to Decisions, the national security community is increasing the focus on technological solutions to analysts’, policymakers’, and warfighters’ overwhelming data challenges.  Solutions often range from cloud-based architectures to machine learning to data mining efforts, and highlight the success of big data analytics in other domains. For instance, in the sports world, Billy Beane’s team of quants out-performed the domain experts, while Nate Silver stuck to his data and proved the political pundits wrong this past election. President Obama’s interdisciplinary analytic team certainly was a contributing factor in his election success. And most pertinent to the national security community, a 2011 Nature article notes, “News Mining Might Have Predicted the Arab Spring.”

Preparing the Environment

Unfortunately, too often these technological solutions occur in a vacuum from domain expertise, and can go terrible astray if in the wrong hands. From being used to justify biases, to producing theoretically incorrect or nonsensical models, to simply ignoring the validity of data, computational analytics should not be viewed as a silver bullet. In complex systems – such as those we are seeing the global environment – the information and data are so noisy that there must be some means to parsimoniously identify the signals. And that is the realm of domain expertise. For instance, social science insights and models can help inform thinking by weeding out irrelevant information, and prioritizing those driving factors behind many of the key challenges in the environment. In this regard, domain expertise can also help make big data relevant, informing computational models with the inclusion of theoretically sound and operationally relevant variables instead of the variable soup that too often populates these complex models that fail to become operationally informative. In addition, with enough data, there will simply be an overwhelming amount of correlations. Domain expertise can weed out the nonsensical correlations, but also enables the exploration of causal mechanisms, while providing context to the findings and interpretation of results. Given the big data environment, the national security community should look for ways to integrate domain experts with the computer scientists and engineers. Computational analytic models and other advanced analytics will not succeed without a domain expert in the loop to help validate the models and vet the data, and ultimately help operationalize the new capabilities.

The dynamic and complex operational environment requires equally dynamic and rigorous capabilities to best inform strategy and operations, add rigor to assessments, and synchronize planning efforts across the globe. In a time of budgetary austerity, great efficiencies can be gained by marrying domain expertise with big data solutions. Big data analytics, when coupled with the insights and context that domain expertise provides, can greatly enhance the national security community’s ability to not only understand the dynamics of the global environment from strategic to tactical levels, but also can help make big data solutions relevant. Given the growing complexities and interdependencies within the operating environment, a smart integration of qualitative expertise and computational models could go a long way in making the big data revolution operationally relevant for the national security community.

____________________

Dr. Andrea Little Limbago is the Chief Social Scientist at Berico Technologies. She has taught courses on international relations, political economy, and development, and spent almost five years as a quantitative social scientist at the Joint Warfare Analysis Center.

Related Articles:

http://www.wired.com/science/discoveries/magazine/16-07/pb_theory

http://www.foreignaffairs.com/articles/139104/kenneth-neil-cukier-and-viktor-mayer-schoenberger/the-rise-of-big-data

http://www.nature.com/news/2011/110913/full/news.2011.532.html

http://www.technologyreview.com/featuredstory/509026/how-obamas-team-used-big-data-to-rally-voters/

http://www.wired.com/insights/2013/05/more-data-more-problems-is-big-data-always-right/

About these ads
3 Comments leave one →
  1. July 25, 2013 3:34 am

    Good way of explaining, and good post to obtain information regarding my presentation subject matter, which
    i am going to convey in institution of higher education.

  2. July 26, 2013 8:32 pm

    Hello there! Would you mind if I share your blog with my facebook group?

    There’s a lot of people that I think would really appreciate your content. Please let me know. Many thanks

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 39 other followers

%d bloggers like this: