Weak Links: The Universal Key to the Stability of Networks and Complex Systems (The Frontiers Collection) (英語) ハードカバー – 2006/3/20
Kindle 端末は必要ありません。無料 Kindle アプリのいずれかをダウンロードすると、スマートフォン、タブレットPCで Kindle 本をお読みいただけます。
"This is an excellent book, which shows the far-reaching consequences of the great variety of weak links. The book has a proper balance between a scientific monograph and a popular approach, and mixes humor with sharp intellect. "Weak Links" is an adventurous, entertainingly eclectic and rich work both for the experts and laymen." (Lászlo-Albert Barabási, Department of Physics, University of Notre Dame, author of the bestseller book, Linked and the 1999 seminal Science paper on the preferential attachment model of scale-free network topology.)
"You have written a very personal, engaging, and unique book that will appeal to readers and get them thinking." (Steve Strogatz, Dept. of Theoretical and Applied Mathematics, Cornell University, author of Sync and the 1998 seminal Nature paper on small worlds.)
"You have done a great service by making the world of networks understandable and clear. I will use your book in my classes." (Caroline S. Wagner, Center of International Science & Technology Policy, George Washington University, author of several science policy-related books including the an upcoming work on the international collaboration in science)
"This book links an exceptionally large number of areas and gives exciting novel information to both the network experts and the science-orinted general readership." (Tamás Vicsek, Dept. Biological Physics, Eötvös University, author of several network-related Nature papers including a method to determine overlapping network modules)
"This masterpiece should serve as an example how science can be discussed. Entertaining yet thought provoking." (György Buzsáki, Board of Governors Professor, Center for Molecular and Behavioral Neuroscience Rutgers University, a leading expert of neuronal networks)
"You have written a true gem of a book; erudite, humane, funny, accessible and thoroughly fascinating. On every page, I find another delight that makes me smile and leads me down new intellectual paths (weak links again!). Thanks to your thorough footnotes, I can delve as deep as I would like into the professional papers. Outstanding - I wish more books were written this way. I have adopted your book as a textbook for my Science of Networks class, and I will recommend it to anyone who ask without hesitation. You did a great service to pedagogy and to this budding science with this magisterial survey. I really appreciate it and my students will, as well." (Daniel J. Bilar, Computer Science Department, Wellesley College MA, USA)
Peter Csermely (50) is a professor at the Semmelweis University in Budapest. A former Fogarty Fellow at Harvard University, his main fields of study are molecular chaperones and networks. In 1996 Dr. Csermely launched a highly successful initiative providing research opportunities for more than 10,000 gifted high school students. He also established the Hungarian National Talent Support Council and the Network of Youth Excellence, www.nyex.info, promoting similar activities in 33 countries. He has published 11 books and more than 200 research papers. Dr. Csermely holds several distinguished appointments including membership of the Wise Persons' Council of the Hungarian President, vice-president of the Hungarian Biochemical Society and has been recipient of numerous international fellowships and awards, for example the 2003 Science Communication Award of the European Molecular Biology Organization and the 2004 Descartes Award of the European Union for Science Communication.
|星5つ 45% (45%)||45%|
|星4つ 35% (35%)||35%|
|星3つ 0% (0%)||0%|
|星2つ 20% (20%)||20%|
|星1つ 0% (0%)||0%|
The author first reviews some basic concepts of networks (basic to those who are well-versed in this, but to others this section would be necessary), and then proceeds in later chapters to introduce his ideas of how weak links influence and stabilize networks, using a variety of research domains as evidence.
However, while being interesting ideas with some merit, the book is written more from the perspective of a mad scientist than from the perspective of a serious researcher... and one that suffers from the "my new theory explains all" syndrome at that! The author acknowledges this, and actually means this to be humorous, but this can be a hit-or-miss. He also has sections where he goes into wild speculations (to his credit, these are well-marked as such, so readers not interested in wild speculations can easily skip these sections). Occasional rantings with Spike (his inner annoying voice) also comes in as a humorous distraction - but again, a hit-or-miss. This does not take away from the merit of the ideas though, it speaks solely to the manner that the book is written, and some readers may enjoy this more than I have.
The main problem that I have is that some of the ideas are not well presented. I was confused about the chaperones in the preface (not a good start), and I find that many of his ideas could be better supported with either more examples from research, or more details about the studies reported. The content should be made more accessible to people without extensive background in the topics covered, especially, as the author likes to point out, researchers from these different fields of research rarely compare notes.
Perhaps a good book to use as a fun introduction to networks with diverse applications, just not exactly what I expected when I ordered it.
Structurally, his book starts with an exposition on network theory and
terminology, then the application and discussion of these concepts to
real-life complex systems on many scales and applied to many domains (physical, natural, technological, social). His main point is, as the reviewer noted above, that 'weak' links (weak: additional/removal does not statistically affect the average of some metric) stabilize systems.
The book has thorough footnotes, one can delve as deep as one would like
into the professional papers. In addition, Csermely is an honest scholar - he shows his hands when there is mere speculation (you have to see the book's unique pictograms to appreciate the effects)
After pouring through several alternatives, I have adopted this book as a
textbook for my Science of Networks class (I'm CS fac at an elite US liberal arts school), and I recommend it to anyone without hesitation for a readable, and learned exposition.
I only have two or three caveats from a specialist's point of view: The
phenomenological discovery of power laws in complex systems is not unusual
and may not be evidence of any SF properties. Scale-free is an abused
term, and I wish the controversy about it were explained a bit more. Also, from a modelling point of view, I wish Doyle and Carlson's work on HOT systems were discussed in more depth.
But these are minor points, relatively speaking. This is a gem of a book:
erudite, humane, funny, accessible and thoroughly fascinating. On every
page, there are delights that lead down new intellectual paths.
Csermely did a great service to pedagogy and to this budding science with
this magisterial survey. Outstanding in its ease of access for intelligent
undergraduates and commendable for intellectual honesty - I wish more
books (textbooks and otherwise) were written this way.
The book is an interesting read if only because its topic matter ranges from network complexity in physical systems, to biological systems, and finally social and cultural systems. Personally I think there are a few longbows drawn, but in fairness Csermely does clearly indicate where he is engaging in speculation. One fascinating discussion was the discourse on pink noise. Pink noise is also known as coloured noise, flicker noise, crackling noise and Barkhausen noise. Seemingly pink noise is present in systems as diverse as solar flares, traffic flows and group decision making, and has a stabilising or relaxing effect. Quoting several scientific sources he postulates that pink noise helps neural synchronisation, which is partly responsible for memory formation. To put it another way if you want to memorise something have Mozart playing in the background rather than bagpipes, because Mozart's music has pink noise properties!
Csermeley's discussion on immunological networks is also interesting. He says an immune system has to solve four problems:
the self/ non-self recognition problem;
the signal to noise problem;
the context problem; and
the response problem.
Now this is interesting because the later three points define the knowledge retrieval problem of a knowledge management system. Apparently weak links are the immune system's mechanism to solve these problems.
A software package typically consists of several hierarchical and modular components, which are bound by strong links. Taking a lesson from the immune system perhaps we need to build software with lots of weak links, and ensure our people and process dimensions also have many weak links? Perhaps these weak links will allow the percolation of knowledge through the human, process, and technology systems. Perhaps our real problem with knowledge management is we try to over-engineer everything and in so doing build strong links rather than weak links. I'm beginning to think weak links matter.
I would recommend this book to anyone who is interested in networks.