[Nets-seminars] Talk next Friday 26th November, Eiko Yoneki (Cambridge) GS/302 16:00

[Richard G. Clegg]

Next Friday our talk is by Eiko from Cambridge.  Also an advanced 
heads-up that we have an extra talk on December 10th from  Vidhyalakshmi 
Karthikeyan and Detlef Nauck of BT -- that will be at 11:30 am.  I will 
send more info nearer the time.

Title: Empirical Approach for Modelling Dynamic Human Contact Networks

Abstract: Increasing numbers of mobile computing devices form dynamic
networks in daily life. In such environments, nodes (e.g. smart phones) are
sparsely distributed and form a time-dependent network. In the EU Haggle
project, we have explored new communication paradigms using dynamic
interconnectedness between people/devices. Efficient forwarding algorithms
for such networks are emerging, mainly based on epidemic routing protocols,
where messages are simply flooded when nodes encounter each other. To reduce
the overhead of epidemic routing, we attempt to uncover a hidden stable
network structure such as a social network, consisting of a group of people
forming socially meaningful relationships such as community. We have proved
that these properties are important aspects of building the forwarding
protocol to improve forwarding efficiency. We have exploited device
connectivity traces from the real world for modelling social network
structure, taking an experimental rather than theoretical approach. The
empirical study of contact networks shares many issues with network-based
epidemiology, and our work has been extended towards understanding the
epidemic spread of infectious diseases. Capturing human interactions will
provide an empirical quantitative measurement of societal mixing patterns to
underpin mathematical models of the spread of close-contact diseases. I will
also briefly introduce our ongoing project 'crowd computing', which combines
mobile devices and social interactions to achieve large-scale distributed
computation. Mobile networks offer substantial aggregate bandwidth and
processing power, and a practical task-farming algorithm in parallel mobile
computing will be investigated by exploiting social structures.


-- 
Richard G. Clegg,
Dept of Elec. Eng.,
University College London
http://www.richardclegg.org/