Job: These CIFRE, Statistical Modeling of Web Buzz Mechanics

Thierry Hamon thierry.hamon at UNIV-PARIS13.FR
Wed May 25 09:02:06 UTC 2011

Date: Mon, 23 May 2011 12:25:37 +0200
From: eustache diemert <ediemert at>
Message-ID: <1306146337.2132.60.camel at gervasuti>

Merci de diffuser à toutes personnes intéressées.

Nous cherchons un candidat pour une thèse CIFRE sur la modélisation
statistique des phénomènes de buzz sur le Web.

PhD Position : Statistical Modeling of Web Buzz Mechanics (Industry

We are looking for a suitably qualified candidate for a challenging
position as an industry funded PhD candidate in statistical modeling
of collective behavior in online communities and social networks. The
successful candidate will work in a team of world class scientists at
BestOfMedia's global R&D center.

An in-depth description of the position will be communicated upon
application to the position.

The Topic :

BestOfMedia R&D teams conduct applied research in the fields of
statistical learning, web mining, information retrieval and natural
language processing. One of the key scientific challenges is to
propose models that accurately describe the mechanics of Web
Buzz. More precisely, the objective is to propose models that explain
the emergence of new topics in news or RSS feeds as well as social
networks and other sources of online information. A second and most
important goal is to be able to rank the potential of a given topic to
create a buzz or burst in interest.

A key area of interest is in detecting, analyzing and modeling the
micro-level features and behavior that impact upon the macro level
performance. In particular, we are interested in being able to profile
different markups that could be detected through NLP techniques and/or
user behavior modeling techniques and help predict the future
performance of new topics.

Current areas of interest include but are not limited to: 

 - Analysis and selection of signal sources
 - Selection of relevant content based features 
 - Selection of relevant user features 
 - Formulating predictive models for information diffusion in large
   scale, online social networks
 - Studying the effect of selected features on information diffusion
   in social networks
 - Designing efficient graph algorithms to find evolving communities
   of interest around topics with correlated change behavior
 - Community detection in dynamic networks
 - Studying and classifying online forum dynamics from a buzz
   mechanics perspective
 - Designing scalable, distributed, real-time algorithms to tackle
   problems in massive dynamic graph datasets

In addition, the successful candidate will collaborate with scientists
and engineers in order to industrialize relevant ideas and
prototypes. He/She will also have the opportunity to analyze and
experiment with the outcome of the applications of his/her work when
put live.

About BestOfMedia:
- Team & Environment
 - young, growing team of world-class scientists with background from
   top academic and industrial labs : EPFL, Max Planck, Yahoo!, etc
 - highly collaborative environment within the team and with external
   companies and labs
 - exceptional working environment in Grenoble, France at the heart of
   the french Alps
 - motivating salary and compensations  

- Data
 - 12 languages
 - ~42+ M unique users per month
 - ~10's of M of web pages
 - very supportive online community  
- Projects
 - automatic key phrases extraction 
 - junk detection for forum content
 - user/content recommender system
 - automatic document tagging
 - dynamic display optimization
 - trends detection
 - buzz prediction
 - social network mining
 - sentiment/opinion mining
 - influence detection

More information about the position and our research interests can be
found by sending an e-mail to ediemert at

Eustache Diemert
Lead Scientist
BestOfMedia Group

4 rue des Méridiens
F-38130 Echirolles

ediemert at
+33 (0)622 744 876

Tom's Labs / BestOfMedia Group

Message diffuse par la liste Langage Naturel <LN at>
Informations, abonnement :
English version       : 
Archives                 :

La liste LN est parrainee par l'ATALA (Association pour le Traitement
Automatique des Langues)
Information et adhesion  :

More information about the Ln mailing list