IoT-AI and Networking 4/7

By Motoyuki Matsunaga, Senior Fellow

 Explanations have been made on the evolution of new concept of IoT-AI Networking in the latest 3 articles in series. References have been made there to the global view on computing combining IoT to connect many things and AI with intelligence ability, the potential of urban system with autonomous factors and the essential points of 3 information evolutions which developed and created them.　 

 In the latter part of this series of articles further explanation will be made on the adaption to urban environment of networking system in polymorphic manner with close association with the facts so far explained in the past articles.

 One of the important features in the adaptive system with polymorphic architecture is the Local Optimization. Local Optimization means the action to pursue the optimization depending on each situation and timing.

 For example, thanks to the invention of LED light it is now possible and getting popular that the lighting is made with the lighting color source to match the human eye perception. Flexible coloring called color rendering properties is now available and one electric light can illuminate with a variety of color and brightness such as daylight, daylight white and bulb color.

 Computer can learn user’s taste and behavior, so, if computers and such lightings are networked, a variety of lighting and illumination can be created to match up with user’s feeling and behavior as computers perceive and recognize what might be optimal for user and will repeat such local optimization for user flexibly depending on the changing environment.

 Environment keeps changing, so computers can learn such change at each time to pursue local optimization.

 In March this year IISE hosted the IoT Study Symposium where Professor Katsuhiro NISHINARI of Research Center for Advanced Science and Technology, the University of Tokyo explained the concept of Quasi-Optimization. If we try to explain the concept of this Quasi-optimization in the case of the above lighting, the concept can be explained that lighting and illumination can be more flexibly produced if the adjustment is made to the situation and change of each room such as study, restroom, and bed-room.

 The lighting (color. brightness) could only be less flexibly produced if the optimized level is decided uniformly with the average overall figures.

 The concept of Quasi-Optimization is the one to be used for the situation and environment like natural world where many factors are ever changing and forecast of the change of environment is difficult.

 In the biological metaphor this Quasi-Optimization concept and the local Optimization by IoT-AI networking concept are close and familiar with each other.

 Many different sensors such as camera, temperature sensors, human sensors are getting connected with network and the word IoT was originated.

 This evolution of Information and Communication Technology (ICT) is often described as computers start to have their own sense organs such as eyes and ears for the real space. Next generation computer network can perceive and learn by themselves combined with AI. Factory machines can be progressed to the next automation stage, and the ability of perception and intelligence will be rapidly developed in 5 years to come.

 A new era is arriving where humans and computers co-exist in the world with mixture of computers, those driven by the order or program predesigned in a conventional manner and those with autonomous perception and learning ability.