Email this Article Email   

CHIPS Articles: Knowing Yourself is key in Cognitive Warfare

Knowing Yourself is key in Cognitive Warfare
Machine learning will enable the naval services to gain organizational insight like never before
By Robert P. Kozloski - January-March 2018
Over two thousand years ago, the Chinese General Sun Tzu famously observed in The Art of War, “Know the enemy and know yourself; in a hundred battles you will never be in peril. If ignorant of both your enemy and yourself, you are certain in every battle to be in peril.”

As an organization, we often take “knowing” ourselves for granted and consider only superficial measures, such as number of ships, readiness levels, and personnel counts. The naval services have an incredible opportunity on the horizon to capture information about the cognitive abilities of our workforce and to convert this information into an operational advantage in information-age warfare.

The incomprehensible amount of raw data generated during World War II stimulated the development of several academic disciplines related to the information and cognitive sciences. Advances in these fields over the past seventy years have provided valuable insights to how humans, computers and organizations process information, and equally as important, to how human limitations affect organizational performance.

Today, military organizations largely treat cognitive skills like their physical counterparts and in doing so; they seek to standardize techniques and performance. This approach is evident in close-order drill, marksmanship, tactics and physical training. While this standardized approach was appropriate for industrial-age models of warfare, where the application of massed fire was key to success in battle, warfare in the information-age promises to be different. Individual and organizational cognitive capabilities will be of paramount importance because of the speed and volume of information available in the modern battlespace.

Before exploring how modern technology can improve cognitive performance, let us briefly draw from several academic observations to understand this issue better.

Herbert A. Simon, a founding father of artificial intelligence, introduced the concept of bounded rationality in the late 1950s in Models of Man: Social and Rational-Mathematical Essays on Rational Human Behavior in a Social Setting. This theory holds that humans are not fully rational and are limited by their cognitive abilities when making decisions. Humans have a finite ability to process information and to explore the results of alternative choices. In Organizations, James G. March and Simon would later apply this theory to organizations and note that because organizations are comprised of humans, human limitations and preferences ultimately affect organizational behavior. Military organizations are subject to the bounded rationality problem, but this constraint often goes unrecognized in practice.

In their seminal work, “Judgment under Uncertainty: Heuristics and Biases,” from 1982, Amos Tversky and Daniel Kahneman similarly observed that humans suffer from limited memory, attention, and the cognitive capacity necessary to identify and process the utility of all options when making decisions. Therefore, humans create mental heuristics (rules of thumb) to make decisions in complex environments. These heuristics rely largely upon previous experience and are subject to predictable errors, or biases. All military officers have unique experiences and personalities that serve as their foundation for cognitive reasoning. It is essential to take these individual differences into consideration when examining future military capabilities. Of note, this pioneering study was funded and directed by the Office of Naval Research, demonstrating the Navy’s enduring interest in this topic.

Certain conditions common to military environments, stress, fear, or fatigue, for example, often reduce cognitive capacity further. In a recent study on the psychological effects of poverty, Johannes Haushofer and Ernst Fehr observed that stress had a negative effect on decision-making and risk perception during controlled experiments. This ultimately created a negative feedback loop, making it difficult for the poorer study participants to improve their long-term situations. This detriment was largely attributed to a cognitive constraint caused by stress. Military officers often encounter different environmental conditions that, if not managed properly, can lead to sub-optimal performance.

These examples illustrate that individual cognitive limitations affect the overall performance of the organization; individuals have unique characteristics that must be taken into account, and factors such as stress cause variation in cognitive performance.

Augmented and virtual reality is an important component of military modeling and simulation training systems. Military leaders, particularly in the naval services, are calling for better systems that permit military personnel to train more frequently — thus creating a virtual gymnasium for developing cognitive skills. The Commandant for the Marine Corps, General Robert Neller, asked recently, “Where’s our Enders Game battle lab kind of thing, where we cannot just give our leadership reps, but we can actually find out who the really good leaders are?” Similarly, the Vice Chief of Naval Operations Admiral Bill Moran called for a holodeck-like training system that contains a “Torpedo room, engine room, bridge room… It knows when you were there last, it knows how effective you were, what your performance levels were, how much experience you have, and it starts to test you.”

Virtual environments should complement, not replace, live training. However, virtual training does offer a significant advantage; it can provide cognitive experience based on environmental conditions that are too difficult, nigh impossible, to replicate in reality. It is for this reason many strategic studies scholars, such as Charles Hill, in Grand Strategies Literature, Statecraft, and World Order and Lawrence Freedman in The Future of War — A History, explore the use of literature, particularly fiction, in developing strategy. Further, recent academic research has demonstrated the effect modern fictional authors, such as Tom Clancy, have had on U.S. national security decision-making.

The transition to virtual (or synthetic) training systems provides an unprecedented opportunity not only to provide unique training experiences; it will also provide a wealth of data on the cognitive processes of military officers at the individual and organizational levels.

Information rich environments as described here are ideally suited for machine learning technology. Virtual training systems that fully incorporate machine learning could provide immediate benefits. What should we do, or leave undone, in a day or a lifetime? How much messiness should we accept? What balance of the new and familiar is the most fulfilling? These may seem like uniquely human quandaries, but they are not. Computers, like us, confront limited space and time, so computer scientists have been grappling with similar problems for decades. And the solutions they’ve found have much to teach us.

In a dazzlingly interdisciplinary work, Brian Christian and Tom Griffiths show how algorithms developed for computers also untangle very human questions. They explain how to have better hunches and when to leave things to chance, how to deal with overwhelming choices and how best to connect with others. From finding a spouse to finding a parking spot, from organizing one’s inbox to peering into the future, Algorithms to Live By transforms the wisdom of computer science into strategies for human living.

Christian and Griffiths observe that military training often suffers from the problem of over-fitting, that is, military personnel train to standardized tasks and routine scenarios and while they become proficient at specific techniques, their ability to respond to new or unforeseen events suffers. Machine learning can prevent this condition from occurring by cross-validating cognitive skills in continually varying information environments.

Further, individual and unit information can be collected and categorized to help gain a deeper understanding of their cognitive abilities and group dynamics (the collective mind of the organization). The data collected in virtual training environments could eventually lead to a military-cognitive activity index (MCAI), similar to the Myers-Briggs Type Indicator (MBTI) personality inventory.

While the MBTI is not without its critics and it is often misused in practice, it does provide a useful framework for understanding the reasons why humans behave differently. An MCAI could provide the same value for understanding how military personnel reason differently. When compared to the MBTI, the MCAI benefits from empirical data collected in virtual training systems, rather than less reliable questionnaire data. Further, data accuracy and reliability would improve if career incentives were tied to actual performance in the training system. In short, military officers would take virtual training more seriously if the results of their cognitive performance, as is their physical performance, were taken into consideration during career decisions, such as command selection or orders to an intellectually demanding duty assignment.

The MBTI is based on four dichotomies that result in sixteen personality types. In my view two variables, uncertainty and complexity, are present in all virtual training systems. The degrees of information uncertainty and environmental complexity must adapt to the performance of the trainee; similar to removing or adding weights to the bar at the gym. In addition to these two variables, I offer four MCAI assessment criteria for consideration:

  • Perception/Interpretation: Military officers should be continually observing their external environment for deviation from expected conditions. This process described in an Office of Naval Research Technical Report, Organizations as Information Processing Systems, can be broken down into three stages: scanning, interpreting, and acting/learning. Virtual environments are ideal for monitoring how individuals preform this task; it will also help to identify individual or organizational biases that may prevent signals from being extracted from noisy environments.
  • Explore-Exploit: When making decisions or planning for military operations, there is always a trade-off for acting on the information at hand or seeking additional information to improve the outcome. Seeking perfect solutions or avoiding risk may affect the point at which military personnel transition from exploring for new information to exploiting the best information available.
  • Logical Preference: Understanding causal relations and mechanisms is critical for the military application of force to achieve a desired outcome successfully. There are several logical or reasoning models to consider: abductive (I suspect most military officers unknowingly align to this pragmatic model), deductive, or inductive, that can be applied to various situations. It is likely that individuals have a predilection for a certain type or combination of models for understanding their environment. Observation in virtual training will help identify how military officers apply actions to outcomes, based on the nature of information available in a given scenario.
  • Networked Learning: A simple framework for understanding useful knowledge holds that there are two types of knowledge: knowledge of “what is” and knowledge of “how to.” Knowledge of both types is stored in documents, publications, on computers, but most importantly, in the minds of others. Modern organizations are founded on the premise of shared cognitive workloads; no one person knows all. An organization must do everything in its power to create both types of knowledge and to facilitate the exchange of knowledge across individuals and teams, without creating unnecessary “noise.” Military officers differ in their ability to access knowledge in a networked environment. This criterion requires individuals to seek out and use the knowledge of others on their team and from vertical and horizontal social networks within their organization and from their external environment.

These four criteria serve as an initial framework for creating a MCAI, and clearly much more research is needed. While the concept is not yet fully developed, it is important to create a vision for modern training systems as it will take time and resources to implement such a vision fully. The discussion heretofore has focused largely on individual-level value afforded by this emerging technology. There are numerous organizational benefits to consider as well.

First, just as sophisticated statistical models, such as those included in sabermetrics, have changed how managers use the talent on their sports teams, identifying cognitive strengths and weaknesses of a staff in real time would allow commanders to manage the cognitive abilities within their organization more effectively. For example, writing for Proceedings, Lt. Brendan Cordial observed, “If the U.S. Navy is going to be led by the best tacticians, officer FitReps must include objective evaluations of warfighting prowess.” This is difficult to accomplish without the use of virtual environments.

Second, advances in computing power would enable the dynamic identification of the cognitive composition of social networks within an organization. For example, how does the performance of an organization comprised of 70 percent of type X and 30 percent type Y personnel differ from an organization with 90 percent of type Y and 10 percent type X? Consider how this type of insight may have helped President John F. Kennedy to identify the biases within his EXCOM during the 1962 Cuban Missile Crisis. Again, working through the different combinations would take significant computing power but it is certainly possible.

Third, the MCAI could be incorporated into a broader heterogeneity index within an organization. Such an index would incorporate a variety of factors such as formal education, pre-military skills and experience, socio-economic factors, gender, race, and even personal interests. The military services already collect most of this information yet it fails to exploit it fully in practice. The development of a heterogeneity index would provide valuable insight for commanders at all levels of the organization. Such a measure would improve problem solving by increasing cognitive diversity and may help prevent group think.

The development of an effective MCAI would directly benefit decision-making and knowledge creation in the near future. It will also aid the adoption of machine-learning into operational systems by helping to understand human-machine teaming. If technological trends hold, the MCAI also will benefit human cognitive augmentation by identifying where and how cognitive performance could benefit from a technological “boost” and how the organization changes practices to accommodate super learners. For example, in collaboration with the Defense Innovation Unit Experimental (DIUx), Naval Special Operations Command initiated a cognitive enhancement project in 2017 with a small group of Special Warfare volunteers at the request of NAVSOC Commander Rear Adm. Tim Szymanski.

The importance of truly “knowing yourself” cannot be understated. Advances in computing technology, particularly machine learning, provide the naval services with the opportunity to know itself like never before. Collecting and analyzing the data generated in virtual environments will enable military organizations to understand the cognitive performance of individuals. As the reliability of information collected in these systems improves, the military services must create a military-cognitive ability index or framework similar to the Myers-Briggs Type Indicator to benefit from this insight. Ultimately, the improved understanding of military cognition will lead to operational advantages in information-age warfare.

Robert P. Kozloski is a Program Analyst with the Department of the Navy temporarily assigned to the University of Pittsburgh. Prior to this position he was a Modeling and Simulation Intelligence Officer at the Defense Intelligence Agency.

The views expressed here are solely those of the author, and do not necessarily reflect those of the Department of the Navy, Department of Defense or the United States government.

Related CHIPS Articles
Related DON CIO News
Related DON CIO Policy

CHIPS is an official U.S. Navy website sponsored by the Department of the Navy (DON) Chief Information Officer, the Department of Defense Enterprise Software Initiative (ESI) and the DON's ESI Software Product Manager Team at Space and Naval Warfare Systems Center Pacific.

Online ISSN 2154-1779; Print ISSN 1047-9988
Hyperlink Disclaimer