BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Computer Laboratory Systems Research Group Seminar
SUMMARY:Entropy Rate of Diffusion Processes on Complex Net
works - Vito Latora (Universita' di Catania\, Ital
y)
DTSTART;TZID=Europe/London:20080807T140000
DTEND;TZID=Europe/London:20080807T150000
UID:TALK13033AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/13033
DESCRIPTION:In the realm of complex networks the concept of en
tropy has been used as a measure to characterize p
roperties of the topology\, such as the degree dis
tribution of a graph. Alternatively\, various auth
ors have studied the entropy associated with ensem
bles of graphs and provided\, via the application
of the maximum entropy principle\, the best predic
tion of network properties subject to the constrai
nts imposed by a given set of observations. The m
ain theoretical and empirical interest in the stud
y of complex networks is in understanding the rela
tions between structure and function. Many of the
interaction dynamics that takes place in social\,
biological and technological systems can be analyz
ed in terms of diffusion processes on top of compl
ex networks\, e.g. data search and routing\, info
rmation and disease spreading. In this talk\, we
show how to associate an entropy rate to a diffusi
on process on a graph. In this context\, the entro
py rate is a quantity more similar to the Kolmogor
ov-Sinai entropy rate of a dynamical system\, than
to the entropy of a statistical ensemble\, and me
asures what is\, on average\, the shortest per ste
p description of the diffusion on the network. The
refore\, a high entropy rate indicates a large ran
domness\, or easiness of propagating from one node
to another\, and can be related to an efficient s
preading over the network Differently from the net
work entropies previously defined\, the entropy ra
te of a diffusion depends both on the dynamical pr
ocess and on the graph topology. This allows us to
use the entropy rate in two different ways: i) to
characterize with a single measure various struct
ural properties of real-world networks\, and ii) t
o design optimal diffusion processes which maximiz
e the entropy. As an example of the powerful possi
bilities of the introduced measure\, we study the
diffusion of random walkers whose motion is biased
on the node degrees.\nJ. Gomez-Gardenes\, V. Lato
ra\, http://xxx.lanl.gov/abs/0712.0278\n\n\nVito L
atora: Dipartimento di Fisica\, Universita' di Cat
ania\, and INFN Italy \n\nhttp://www.ct.infn.it/~l
atora/bio.html
LOCATION:FW26\, Computer Laboratory\, William Gates Builidi
ng
CONTACT:Eiko Yoneki
END:VEVENT
END:VCALENDAR