Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Advances in applied self-organizing systems (2nd ed.)
Prokopenko M., Springer Publishing Company, Incorporated, New York, NY, 2013. 436 pp. Type: Book (978-1-447151-12-8)
Date Reviewed: Feb 4 2014

Truly distributed information processing is one of Mother Nature’s best kept secrets. But we have been trying to make it happen. Before 1950, the ripple-carry adder processed (local) patches of (global) numbers in a synchronous array of bit, decimal, or octal units. That accomplishment, the local processing of parts of global data structures, was a step in the right direction. In the 1950s, on the other hand, von Neumann and Ulam gave us cellular automata, as if that had something to do with the mystery of biological information processing! These were arrays that processed global patterns locally, but synchronously and in a highly constrained architecture. In the 1970s, John McCarthy was making a serious effort to understand the logic of distributed processes acting on local patches of global structures. But there was always synchronization, a biologically inappropriate crutch implying the existence of a global clock. Synchronization, broadcast communication, and globally shared memory do not occur in nature.

A precisely specified architecture may not have been a violation of our desire to distribute everything, but it was definitely not natural. The necessary insight did not come easily; decades later, a keynote speaker in an international supercomputing conference (Orlando, 1988) professed distributed control to be unimaginable, even with the crutches of synchronization and prescribed architecture. In 1996, Abelson (MIT) [1] and others began promoting “amorphous” computing, which is amorphous in that everything at the global level is unstructured and distributed, activity is asynchronous, and the architecture is not prescribed. The MIT group had examples inspired by Turing [2], Ashby [3], and modern military intelligence networks.

Now, a decade into the 21st century, there are dozens of journals and conferences devoted to self-organization, amorphous computing, morphogenesis, and distributed computing (most published by Springer). Talk of truly distributed information processing systems, sophisticated enough to support self-organizing behavior (in other words, morphogenesis in Turing’s broad definition of the word), is in the air worldwide.

Last year, Mikhail Prokopenko put together this collection of 15 chapters by 25 established researchers. The excellent introduction is followed by these chapters: “Design Versus Self-Organization”; “Foundations and Formalizations”; “Self-Organizing Traffic Lights”; “Self-Organizing Sensing Structures”; “Decentralized Decision Making”; “Learning Mutation Strategies”; “Self-Organization as Phase Transition--A Study Based on Entropy”; “Microscopic Robots in Medical Applications”; “Songline Processors”; “Self-Organizing Nomadic Services”; “Immune System”; “Formal Immune Networks”; “Self-Organizing Visualization”; “Memristive Excitable Automata”; and “A Turing Test for Emergence.”

Among other topics in these chapters, the authors discuss the use of an appropriate (Shannon) entropy to measure organization, which is the heart of Schrödinger’s What is life? [4] and one of the few concepts certain to appear in any successful future theory of self-organizing systems. Given a dynamical system with state-transition space [S,→] and event-space EP(S), every probability measure pt:E→ [0,1] suggests a notion of organization that is formalized and measured by the associated entropy

The paths from a given initial state s0 can be represented by a sequence of measures p1 → … pt+1 → … indicating likely next states. The system is self-organizing if, for some T, Hpt > Hpt+1 for all t>T. This representation is tractable enough to support proofs of self-organizing behavior and even the construction of individual programs, though this is beyond the present text.

A book of this sort on a developing science may be expected to have most of its takeaway value in the introductions and conclusions of its chapters. The meat of a chapter will be of value to the reader whose work is very close to that of the authors. In this sense, there is much of interest in Prokopenko’s collection.

Reviewer:  W. Richard Stark Review #: CR141964 (1405-0309)
1) Abelson, H. et al. Amorphous computing. AI Memo 1665, (1999), 1–20. http://groups.csail.mit.edu/mac/projects/amorphous/papers/aim1665.pdf.
2) Turing, A. M. The chemical basis of morphogenesis. Phil. Trans. Royal Society of London B 237, 641 (1952), 37 – 72. http://dx.doi.org/10.1098/rstb.1952.0012
3) Ashby, W. R. Principles of self-organizing dynamic systems. Journal of General Psychology 37 (1947), 125 – 128.
4) Schrödinger, E. What is life? (reprint). Cambridge University Press, New York, NY, 2012.
Bookmark and Share
 
Self-Modifying Machines (F.1.1 ... )
 
 
Connectionism And Neural Nets (I.2.6 ... )
 
Would you recommend this review?
yes
no
Other reviews under "Self-Modifying Machines": Date
Complex systems dynamics
Weisbuch G., Ryckebushe S. (trans.), Addison-Wesley Longman Publishing Co., Inc., Boston, MA, 1991. Type: Book (9780201528879)
Dec 1 1991
On the computational efficiency of symmetric neural networks
Wiedermann J. Theoretical Computer Science 80(2): 337-345, 1991. Type: Article
Mar 1 1992
Introduction to the theory of neural computation
Hertz J., Krogh A., Palmer R., Addison-Wesley Longman Publishing Co., Inc., Boston, MA, 1991. Type: Book (9780201503951)
Jan 1 1993
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy