[Hackmeeting] [ROMA] - Tor Vergata - Seminario 13 Luglio su …

Nachricht löschen

Nachricht beantworten
Autor: ZioPRoTo (Saverio Proto)
Datum:  
To: hackmeeting
Betreff: [Hackmeeting] [ROMA] - Tor Vergata - Seminario 13 Luglio su Large Scale Social Networks
Salve,

Per tutti gli interessati, il giorno Martedi' 13 alle ore 14.30 in aula
R3 della facoltà di Ingegneria a Roma Tor Vergata (Nuovi Edifici) si
svolgerà un seminario dal titolo "Analyzing Large-Scale Social
Networks from Data to Applications".
Il seminario sarà tenuto da Alessandra Sala, ricercatrice Post-Doc
presso l'Università di Santa Barbara.
Segue Abstract del seminario e breve biografia.

Saverio

------------------------------------

*Analyzing Large-Scale Social Networks from Data to Applications*

Large scale social networks pose a number of interesting and novel research
problems, from the pure characterization of the structure  of the
network data to
challenging algorithmic issues in designing  new social services and
infrastructures.  Today's researchers rely on realistic social network
datasets to
evaluate their systems and applications.

Due to the time and cost to gather real datasets via direct measurements,
researchers are anonymizing and sharing a small number of valuable datasets
with the community.  Unfortunately, there are a number of obstacles that
limit
researcher in sharing data: concerns about de-anonymization attacks to
reveal
private information, increasing costs of distributing large datasets,
and the limited
statistical confidence derived by experimental from  a small number of
available
social graphs.  The use of measurement-calibrated graph models is an
attractive
alternative to sharing datasets. Researchers could use graph models to
generate
synthetic graphs statistically similar to the original graph; however,
selecting a
graph model able to accurately reproduce the complex structures of
online social
networks is a non trivial problem.  We use both graph metrics and
application level
benchmarks to evaluate several graph models using as input several real
Facebook datasets ranging from 30,000 to 3 million edges.  Our results
show that
there are two graph models, dK and Nearest Neighbor, that have the
potential to
provide realistic synthetic social graphs.



*Short biography*

She is a Researcher in the Department of Computer Science at University of
California, Santa Barbara, USA. From May to 2008 to May 2010 she worked at
University California Santa Barbara as Post Doctoral Scholar affiliated
with the
Current Lab research group led by Prof. Ben Y Zhao. She received her PhD
(2008)
and Laurea Degree (2004) cum Laude from the department of  Computer Science
at the Università degli Studi di Salerno, Italy.
Her research focus lies in Distributed Systems with an emphasis on
communication
and privacy issues in  large scale systems. Her main interest is to
develop efficient
distributed systems that allow robust and flexible services as  scalable
search or
resource sharing, efficient data dissemination, or reliable anonymous
communication.  She is passionate in designing algorithms and systems with
strong theoretical foundations, and in discovering mathematical models
which
govern those networks.
Per accedere agli archivi, modificare le proprie preferenze o cancellarsi dalla lista:
https://lists.uniroma2.it/index.html/info/fondamenti-internet