Heikki Hyötyniemi, Dr.Tech, was a professor of automation technology at Helsinki University of Technology from 2001 until his early retirement in 2009. An author of some 200 publications, his early research interests included artificial neural networks and methods of artificial intelligence in general; later, his studies have expanded into the fields of natural philosophy, trying to understand the nature of complex (living) systems in general.
Hyötyniemi's book "Enformaatioteoria - elämänvoiman perusteet" (now available on Suomalainen Kirjakauppa or Akateeminen Kirjakauppa) explains how process philosophical principles can open up fresh possibilities when modeling complex systems. His findings are summarized in English in this HIIT Seminar.
Abstract:
Enformation is "energetic information": being a form of energy, it has the capacity to make changes in the world. In this way, it defines "natural semantics" that gives meaning to emergent structures (but no purpose). Following the enformation flow it turns out that the diversity in natural systems, for example, can be explained in a novel way in terms of new kinds of "emergent ecolockers".
Technically speaking, enformation can be seen as (co)variation, so that the resulting optimal structures can be interpreted in terms of principal subspace analysis (thus making it possible to implement practical machine learning algorithms). The simplicity of the structures makes it possible to efficiently analyze their properties: it turns out that maximization of enformation on the low level changes to minimization as seen from above. The emergent cost criteria make it possible to find connections to classical "least action" related principles.
Further, it turns out that the emergent systems constitute elementary control loops: coupling of the systems changes their environment, "equalizing" the variables. And the control paradigm helps to understand system evolution: better controls consistently increase "subjective entropy". It can be said that the controls change into interlocked "monads", or "whirls" in the enformation flow; these can be seen as atomary structures in complex systems.
Studying the control system dynamics closer, it turns out that the systems constitute elementary resonators, making it possible to escape from time domain to frequency domain. In this way, it becomes possible to understand interactions among systems: the systems are seen as filters and the signals as spectra. It is fields that make it possible to reach a yet higher level of emergence - the "directed antennas" perhaps explaining even the essence of matter and mind.
http://neocybernetics.com/enformaatioteoria-kirja/
Last updated on 29 Apr 2013 by Antti Ukkonen - Page created on 29 Apr 2013 by Antti Ukkonen