Page 263 - Artificial Intelligence for the Internet of Everything
P. 263

CHAPTER 13

              Compositional Models

              for Complex Systems


              Spencer Breiner*, Ram D. Sriram*, Eswaran Subrahmanian* ,†
              *
              Information Technology Lab, National Institute of Standards and Technology, Gaithersburg, MD,
              United States
              †
              Institute for Complex Engineered Systems, Carnegie Mellon University, Engineering and Public Policy,
              Pittsburgh, PA, United States

              13.1 INTRODUCTION

              In today’s world, computation is reaching into everyday life at home and at
              work, driven by two major trends: artificial intelligence (AI) and the Internet
              of Things (IoT). We intend both terms to have a broad sense, AI to encom-
              pass more-or-less all data processing and IoT to cover any sensing or actu-
              ation in the physical world. Though in different ways, both of these offer
              some possibility for the substitution of human labor; AI for mental tasks
              and the IoT for physical ones. Of course we will be interested mostly in
              the augmentation of human performance, rather than its replacement, but
              either way these new capabilities invite us to rethink the design of systems
              and processes in nearly every domain.
                 The emergence of this vision for the future has raised a thicket of new
              trends and buzzwords: IoT, the Internet of everything, cyber-physical sys-
              tems, sociotechnical systems, smart systems, complex systems, and systems of
              systems. The welter of terminology obscures the fact that all of these areas
              face the same fundamental problems regarding system representation, anal-
              ysis, and synthesis.
                 This is not to suggest that there is no difference between these fields. Those
              buildingtheIoTarenecessarilyconcernedwithflakynetworksandlow-power
              computing; cyber-physical systems need not be sociotechnical, and vice versa.
              However, we believe that the unique features of these fields obscure the most
              important issues at play and are outweighed by the an overall similarity. More-
              over, many of the systems of interest cut across these artificial boundaries; an
              airplane checks most of the boxes above, and a hospital touches all of them.
                 With new promise comes new risk, and as we integrate these technol-
              ogies into critical systems like health care, manufacturing, banking, and


              Artificial Intelligence for the Internet of Everything  2019 Published by Elsevier Inc.
              https://doi.org/10.1016/B978-0-12-817636-8.00013-2  All rights reserved.  241
   258   259   260   261   262   263   264   265   266   267   268