Workshop - "Complexity in Online Social Networks and Big Data", 18 September 2013
Date: September 18, 2013
1. Venue and structure
“Complexity in Online Social Networks and Big Data” is a Satellite Workshop of the European Conference on Complex Systems ECCS 2013 hosted in Barcelona. The workshop will take place in September 18 and will encompass 5 invited talks of one hour each and a poster session including short presentations for each poster.
2. Workshop description
The rapid proliferation of Online Social Networking (OSN) services like Facebook, Twitter and Google+ has made a profound impact on the Internet and tends to reshape its structure, design, and utility. Industry experts believe that OSNs create a potentially transformational change in consumer behavior and will bring a far-reaching influence on traditional industries of content, media, and communications.
The main goal of this workshop, which is mainly related to Conference Tracks of “Social Systems” and “Information and Communication Technologies” is to serve as a focal point for exchanging state-of-the-art results and innovative ideas on how to address the problems and opportunities related with Complexity in Online Social Networks and Big Data. Emphasis will be given on modeling the dynamics of online social networking services from the viewpoint of the Science of Complex Networks and exploiting theoretical insights for the design, modeling and simulation of massive-scale DOSN infrastructures.
3. Invited speakers
(Telefonica TID Research)
|08:45 - 09:00||Welcome by Asst. Prof. Sarunas Girdzijauskas (Royal Institute of Technology (KTH), Sweden)|
|Morning session (chair: Prof. Marián Boguñá (University of Barcelona, Spain) )|
|09:00 - 10:00||Keynote 1: Online and offline social networks, by Alain Barrat (C.N.R.S, France) abstract, presentation|
|10:00 - 11:00||Keynote 2: Social influence and recurrent mobility underlie background fluctuations of electoral processes, by Maxi San Miguel (IFISC, Spain), abstract, presentation|
|11:00 - 11:30||Coffee Break|
|11:30 - 12:30||Keynote 3: A tale of 2 continents and 4 cities about the influence of demographics and social constraints on ride-sharing, by Nikolaos Laoutaris (Telefonica TID Research, Spain),abstract, presentation|
|12:30 - 14:00||Lunch|
|Noon session (chair: Prof. Elena Ferrari (University of Insubria, Italy))|
|14:00 - 15:00||Keynote 4: Challenges and Opportunities of Context-Aware Computing, by Pankaj Mehra,(Fusion-io, USA), abstract|
|15:00 - 16:00||Keynote 5: Mining Propagation Data (in Social Networks), by Francesco Bonchi, (Yahoo! Research, Spain), abstract, presentation|
|Afternoon session (chair: Prof. Fragopoulou Paraskevi (Foundation for Research and Technology - Hellas (FORTH), Greece))|
|16:00 - 16:30||Poster Presentations|
|16:30 - 17:00||Coffe Break
|17:00 - 18:30||Poster Session|
5. Organizing committee
- Benjamin Mandler (IBM Haifa Labs, Israel)
- Elena Ferrari (University of Insubria, Italy)
- Eliezer Dekel (IBM Haifa Labs, Israel)
- Fragopoulou Paraskevi (Foundation for Research and Technology - Hellas (FORTH), Greece)
- George Pallis (University of Cyprus, Cyprus)
- M. Ángeles Serrano (University of Barcelona, Spain)
- Marián Boguñá (University of Barcelona, Spain)
- Marios D. Dikaiakos (University of Cyprus, Cyprus)
- Pietro Lio (University of Cambridge, UK)
- Sameh El-Ansary (Peerialism AB, Sweden)
- Sarunas Girdzijauskas (Royal Institute of Technology (KTH), Sweden)
- Seif Haridi (Royal Institute of Technology (KTH), Sweden)
- Sotiris Ioannidis (Foundation for Research and Technology - Hellas (FORTH), Greece)
- Yasmin Merali (wbs, UK)
The workshop will be organized in the frame of the European Marie Curie Initial Training Networks “iSocial” project and the RECOGNITION grant 257756, an EC - FP7 Future Emerging Technologies project.
New technologies give new ways of studying social and behavioral networks thanks to unprecedented access to different types of datasets. In this talk, I will first present recent work on a detailed empirical study of an interest-based online social network, showing the presence of strong heterogeneities between actors, and how the presence of homophily can be measured. I will also show how longitudinal studies allow us to show that link formation is driven by selection mechanisms and that links are the substrate of influence effects. Influence is measured both through the alignment of users' profiles and by tracking the patterns of diffusion of specific items across users' profiles. Finally, I will consider the issue of comparing the behavior of individuals in online social networks and in offline behavioral networks of face-to-face proximity. I will show, in the case of the network of attendees of a scientific conference, that the existence of an on-line link between individuals is strongly correlated with the strength of their face-to-face presence, as well as with the similarity of their social contacts in physical space. Based on these findings, I will characterize how accurately the existence of an on-line social link can be predicted by using the measured properties of face-to-face presence.
Social influence among individuals is at the core of collective social phenomena such as the diffusion of innovations, social learning, the dissemination of ideas, beliefs or behaviors. Despite the different mechanisms proposed to implement inter-agent influence in social models, a satisfactory confrontation between data and model predictions has not been reached yet. Here we advance in this direction introducing a model that reproduces generic features observed in vote-shares of the US presidential elections at different geographical levels. Our approach incorporates spatial and population diversity as input data into an opinion model in which individuals’ mobility provides a proxy for social context, while peer imitation implements social influence. Our results account for the observed stationary background fluctuations in voting behavior, that is, the dispersion of the vote-shares across counties, congressional districts and states. They also account for spatial correlations decaying logarithmically with spatial distance. We offer a general framework for the micro-macro connection between the individual mechanisms of interaction and society-wide patterns of collective behavior in opinion formation.
Ride-sharing on the daily home-work-home commute can help individuals save on gasoline and other car-related costs, while at the same time reduce traffic and pollution. This paper presents a quantitative study of the potential of ride-sharing in two European and two US cities based on mobility data extracted from 3G Call Description Records (CDRs) and Online Social Networks (OSN) geo-tagging. We start with ride-sharing among neighbors for which we derive up- per bounds based on the assumption that commutes occur synchronously and that any passenger can be matched with any driver. We show that a small detour distance of 0.8 Km leads to an impressive 55% reduction of traffic. This value reduces to 26% for asynchronous commutes and a pick-up tolerance of 10 minutes. The latter result can be improved to 47% by permitting additional passengers to be picked-up en-route using e.g., smartphone apps. We then introduce social constraints on the relationship between drivers and passengers. Riding only with immediate friends derived from calling patterns drops the potential of ride-sharing to less than 1%. Riding with OSN friends leads to a better performance of 8% explained by the fact that OSNs friend sets are typically larger. Riding with "friends-of-friends" causes traffic reduction to increase to : 6% based on CDRs and 23% based on OSNs. Last, we study the impact of demographics and population density on the potential of ride-sharing in the four cities of our study.
Various enterprises are looking to collect and manage vast amounts of information about their users. Some seek to acquire particular dimensions of the larger human context, such as social or location. Yet others seek to integrate the entire context. In this talk, I will describe the architecture and economics of context data infrastructures, from the outside in. I will give concrete examples of techniques used in extracting, indexing, and querying context data. We will look at the mechanics and economics of contextualization, by examining the whys, wherefores, and hows of it.
With the success of online social networks and microblogging platforms such as Facebook, Flickr and Twitter, the phenomenon of influence-driven propagations, has recently attracted the interest of computer scientists, information technologists, and marketing specialists. In this talk we take a data mining perspective and we discuss what (and how) can be learned from a social network and a database of traces of past propagations over the social network. Starting from one of the key problems in this area, i.e. the identification of influential users, by targeting whom certain desirable marketing outcomes can be achieved, we provide an overview of some recent progresses in this area and discuss some open problems.
Gossip-based Partitioning and Replication
DOSN Reference Architecture, Challenges and Opportunities
Large Scale Cross-Document Coreference Resolution
(poster available in:pdf)
Decentralised Social Networking Platforms: Current Status and Trends
Modelling and Simulation in Decentralised Online Social Networks
Security Issues in DOSNs
Data Placement, Replication, Distribution and Streaming Services
A peer-to-peer decentralized protocol for k-Leader Election
On HTTP Live Streaming in Large Enterprises
Community Based Identity Management onDecentralized Social Networks
(poster available in: pdf)
Structure and Evolution of Social Networks
8. Photos of the Workshop