Date : | From 2011-03-01 To 2011-04-01 |

Advisory committee : | |

Local coordinators : | Bing-Hong Wang, Ke Xu, Haijun Zhou(Contact Person),Tao Zhou |

International coordinators : | Mikko Alava, Erik Aurell, Yi-Cheng Zhang |

This five-week program (28 Feb-01 April, 2011) is a continuation of the 2008 KITPC program “Collective Dynamics in Information Systems”. the first two weeks focus on topics (1) and (2),and the last two weeks focus on topics (4) and (5), and the middle week will discuss issues that are interesting to both the spin-glass/information science community and complex network community. The main focuses of this new program are:

(1) Statistical Physics of constraint satisfaction problems and learning problems

Methods from statistical physics are gaining increasing acceptance in information theory and computer science [see, e.g., the recent book by Mezard and Montanari, Information, Physics, and Computation, Oxford University Press, 2009]. In the last few years, statistical physicists in the field have increasingly moved beyond inference to learning. Technically, inference means finding out properties of a model, which is defined and known. Learning means finding out what the model is from observations. Arguably this has an even wider applicability to computer and information sciences. New statistical physics-inspired algorithms has the potential to revolutionize neural-science (learning from measurement on many neurons at the same time) and 'omics' (learning from measurements on many genes at the same time), and certainly also many other fields of science and technology.

(2) Statistical physics of Peer-to-peer (P2P) systems

Computational and communicating entities are spreading over the face of the earth. PCs and cell phones, alive and connected at any one time, already count in the billions or tens of billions. With the projected exploration of sensor networks, embedded systems and various forms of "smart dust", the total aggregate number will soon be in the trillions. Only a very small fraction of this computational and communicative power is used to execute centrally designated and coordinated tasks. Most entities lead partly autonomous lives, executing locally stored programs, and reacting to local information and input. The description of large collections of such entities, and their properties in the large, is a new promising area of statistical physics.

(3) Network-based information physics

The research fields of information physics are widespread. All researches on information based on internet and carried on by means of physics could be called information physics. The important questions that will be discussed in this program include: (a) Data mining and presonalized recommendation on world-wide web; (b) User bahavior analysis in Web; (c) Internet and WWW's frames and mechanism of evolution; and (d) Dynamics research on Internet and WWW.

(4) Statistical physics of complex networks

http://tfy.tkk.fi/~mja/

http://www.csc.kth.se/forskning/cb/cbp/homepages/eaurell/kth_homepage.html

http://www.maia.ub.es/~maneva/

http://lptms.u-psud.fr/membres/mezard/

http://www2.warwick.ac.uk/fac/sci/maths/people/staff/amin_coja_oghlan/

http://www-adsys.sys.i.kyoto-u.ac.jp/tt/Tanaka-e.html

http://people.epfl.ch/rudiger.urbanke

http://power.itp.ac.cn/~zhouhj/