Page 19 - Artificial Intelligence for the Internet of Everything

6 Artificial Intelligence for the Internet of Everything

Adam Fouse, and Robert McCormack. Levchuk is a Senior Principal
Research Scientist and Corporate Fellow at Aptima Inc. in Woburn,
MA, and is an expert in relational data mining, distributed inference, and
reasoning under uncertainty. Professor Pattipati is a Board of Trustees Dis-
tinguished Professor and a UTC Professor in Systems Engineering in the
Department of Electrical and Computer Engineering at the University of
Connecticut; Serfaty is the Chief Executive Officer, Principal Founder,
and Chairman of the Board of Directors of Aptima; Fouse is the Director
of the Performance Augmentation Systems Division and Senior Research
Engineer with Aptima; and McCormack is the Principal Mathematician
and Lead of Team and Organizational Performance at Aptima. From the
perspective of the authors, for modern civilization, IoT, ranging from health
care to the control of home systems, is already integral in day-to-day life.
The authors expect these technologies to become smarter, to autonomously
reason, to act, and to communicate with other intelligent systems in the
environment to achieve shared goals. The authors believe that sharing the
goals and tasks among IoT devices requires modeling and controlling these
entities as “teams of agents.” To realize the full potential of these systems, the
authors argue that scientists need to better understand the mechanisms that
allow teams of agents to operate effectively in a complex, ever-changing, but
also uncertain future. After defining the terms they use in their chapter, the
authors construct the framework of an energy perspective for a team from
which they postulate that optimal multiagent systems can best achieve adap-
tive behaviors by minimizing a team’s free energy, where energy minimi-
zation consists of incremental observation, perception, and control phases.
Then the authors craft a mechanism with their model for the distribution
of decisions jointly made by a team, providing the associated mathematical
abstractions and computational mechanisms. Afterwards, they test their ideas
experimentally to conclude that energy-based agent teams outperform
utility-based teams. They discuss different means for adaptation and scales,
explain agent interdependencies produced by energy-based modeling, and
look at the role of learning in the adaptation process. The authors hypoth-
esize that to operate efficiently in uncertain and changing environments, IoT
devices must not only be sufficiently intelligent to perceive and act locally,
but also possess team-level adaptation skills. They propose that these skills
must embody energy-minimizing mechanisms that can be locally defined
without the need for agents to know all global team-level objectives or con-
straints. Their approach includes a decomposition of distributed decisions.
The authors introduce the concept of free energy for teams and the