Social-Behavioral Modeling for Complex Systems

Paul K. Davis, Angela O'Mahony, Jonathan Pfautz, Paul K. Davis, Angela O'Mahony, Jonathan Pfautz

• Publisher: Wiley
• Year: 2019
• Print ISBN: 9781119484967
• eBook ISBN: 9781119484974
• Edition: 1
• Topic: Technology & Engineering
• Subtopic: Electrical Engineering & Telecommunications
• Index: Technology & Engineering

This volume describes frontiers in social-behavioral modeling for contexts as diverse as national security, health, and on-line social gaming. Recent scientific and technological advances have created exciting opportunities for such improvements. However, the book also identifies crucial scientific, ethical, and cultural challenges to be met if social-behavioral modeling is to achieve its potential. Doing so will require new methods, data sources, and technology. The volume discusses these, including those needed to achieve and maintain high standards of ethics and privacy. The result should be a new generation of modeling that will advance science and, separately, aid decision-making on major social and security-related subjects despite the myriad uncertainties and complexities of social phenomena. 

Intended to be relatively comprehensive in scope, the volume balances theory-driven, data-driven, and hybrid approaches. The latter may be rapidly iterative, as when artificial-intelligence methods are coupled with theory-driven insights to build models that are sound, comprehensible and usable in new situations.

With the intent of being a milestone document that sketches a research agenda for the next decade, the volume draws on the wisdom, ideas and suggestions of many noted researchers who draw in turn from anthropology, communications, complexity science, computer science, defense planning, economics, engineering, health systems, medicine, neuroscience, physics, political science, psychology, public policy and sociology. 

In brief, the volume discusses:

• Cutting-edge challenges and opportunities in modeling for social and behavioral science
• Special requirements for achieving high standards of privacy and ethics 
• New approaches for developing theory while exploiting both empirical and computational data
• Issues of reproducibility, communication, explanation, and validation
• Special requirements for models intended to inform decision making about complex social systems

Part I
Introduction and Agenda

1
Understanding and Improving the Human Condition: A Vision of the Future for Social‐Behavioral Modeling

Jonathan Pfautz¹, Paul K. Davis² and Angela O'Mahony²

¹ Information Innovation Office (I20), Defense Advanced Research Projects Agency, Arlington, VA, 22203‐2114, USA

² RAND Corporation and Pardee RAND Graduate School, Santa Monica, CA, 90407-2138, USA

Technology is transforming the human condition at an ever‐increasing pace. New technologies emerge and dramatically change our daily lives in months rather than years. Yet, key aspects of the human condition – our consciousness, personalities and emotions, beliefs and attitudes, perceptions, decisions and behaviors, and social relationships – have long resisted description in terms of scientific, falsifiable laws like those found in the natural sciences. Past advances in our knowledge of the human condition have had valuable impacts, ¹ but much more is possible. New technologies are providing extraordinary opportunity for gaining deeper understanding and, significantly, for using that understanding to help realize the immense positive potential of the humankind.

In the information age our understanding of the human condition is deepening with new ways to observe, experiment, and understand behavior. These range from, say, identifying financial and spatiotemporal data that correlate with individual well‐being to drawing on the narratives of social media and other communications to infer population‐wide beliefs, norms, and biases. An unprecedented volume of data is available, an astonishing proportion of which describes human activity and can help us explore the factors that drive behavior. Statistical correlations from such data are already helping to inform our understanding of human behavior. New experimentation platforms have the potential to support both theory‐informed and data‐driven analysis to discover and test the mechanisms that underlie human behavior. For example, millions of users of a social website or millions of players of online games can be exposed to different carefully controlled situations – within seconds – across regional and cultural boundaries. Such technologies enable heretofore impossible forms and scales of experimentation. At the same time, these new capabilities raise important issues of how to perform such experimentation, how to correctly interpret the results, and, critically, how to ensure the highest ethical standards.

Such advances mean that theory development and testing in the social‐behavioral sciences are poised for revolutionary changes. Behavioral theories, whether based on observation, in situ experiments, or laboratory experiments can now be revisited with new technology‐enabled instruments. Applying these new instruments requires confronting issues of reproducibility, generalizability, and falsifiability. Doing so will help catalyze new standards for scientific meaning in the social‐behavioral sciences. The massive scale of some such studies will require complex experimental designs, but these could also enable substantially automated methods that can address many problems of reproducibility and generalizability.

Similarly, representation of knowledge about the human condition is poised for revolution. Using mathematics and computation to formally describe human behavior is not new (Luce et al. 1963), but new and large‐scale data collection methods require us to reconsider how to best represent, verify, and validate knowledge in the social and behavioral sciences. New approaches are needed to capture the complex, multiresolution, and multifaceted nature of the human condition as studied with different observational and experimental instruments. Capturing this knowledge will require new thinking about mathematical and computational formalisms and methods, as well as attention to such engineering hurdles as achieving computational tractability.

Advances in knowledge representation will also motivate advances in social and behavioral science. The need for accuracy and precision in describing the current understanding of the human condition will require models with structured descriptions of data sources, data interpretations, and related assumptions. These will support calibration, testing, and integration and, critically, identifying gaps in current theory and instruments. Mathematical, computational, and structured qualitative can models provide a comprehensive epistemology for the social and behavioral sciences – describing not only what is known but also the certainty and generalizability of that knowledge. This will allow comparing different and even conflicting sources of knowledge and resulting theories, as well as identifying future research needs. The combination of computational models with new technology‐enabled instruments for studying human behavior should result in a tightly coupled and partly automated ecosystem that spans data collection, data analysis, theory development, experimentation, model instantiation, and model validation. Such an ecosystem, if constructed to maximize the soundness of the science, would radically transform how we pursue our understanding of the human condition.

Revolutions in the scientific process of creating and encoding knowledge about human behavior will allow applications that aid human decision‐making. Computational models provide the means to readily apply (and democratize nonexpert access to) knowledge of human behavior. However, access to an ever‐expanding body of knowledge about the human condition must be appropriately managed, especially as techniques for reflecting and combining the inherent uncertainties in the growing knowledge base are developed. Also, we must understand how to use these increasingly accurate models and how to quantify and share information transparently, including information about uncertainty, so as actually to assist human decisionmaking rather than increase confusion. Accurately representing the growing bounds of our knowledge about the human condition is essential for ensuring ethical application of the knowledge and maximizing its benefit for society (Muller 2018).

Challenges

We are at the beginning of an era in which sound computational models of human behavior and its causes can be constructed. Such models have vast potential to positively affect the human condition. Yet, despite this heady promise, a great deal of creativity and innovation will be needed to surmount the considerable challenges. Even describing these challenges remains a subject for scholarly debate; therefore, the list of challenges below should be considered representative rather than comprehensive (see also Chapter 2 , Davis and O'Mahony 2018).

Challenge One: The Complexity of Human Issues

First and foremost, it is for good reasons that the social and behavioral sciences have not progressed in the same way as the natural sciences. The human condition is inherently complex with overlapping multiresolution features across multiple dimensions (e.g. from short to long time frames, from individual behavior to group activities and governance, and from individual neuronal physiology to brain‐region activation, psychophysical action, cognitive task performance, and intelligence and consciousness). Further, the...

Table of contents

• Cover
• Table of Contents
• Foreword
• List of Contributors
• About the Editors
• About the Companion Website
• Part I: Introduction and Agenda
• Part II: Foundations of Social-Behavioral Science
• Part III: Informing Models with Theory and Data
• Part IV: Innovations in Modeling
• Part V: Models for Decision-Makers
• Index
• End User License Agreement