It’s my pleasure to announce that our paper “Graph-based Sybil detection in social and information systems” has received the Best Paper Award in its track (there were a total of 54 full papers). The award also includes free conference registration for next year.
The IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM) is one of the top conferences in the field (13% acceptance rate this year). The conference provides an interdisciplinary venue that brings together practitioners and researchers from a variety of SNAM fields to promote collaborations and exchange of ideas and practices. ASONAM 2013 is intended to address important aspects with a specific focus on the emerging trends and industry needs associated with social networking analysis and mining. The conference solicits experimental and theoretical works on social network analysis and mining along with their application to real life situations.
Starting this Sept, I will be joining Telefonica Research as a research intern. I will be working with Dionysios Logothetis and other Telefonica researchers in the Distributed Systems Group. My internship will focus on the system side of big graph mining. In particular, we are looking into designing new algorithms for incremental graph processing in Bulk Synchronous Parallel (BSP) computation model. To evaluate our approach, we plan to extend Apache Giraph and use Sybil detection in OSNs as a representative application.
Our latest research on Graph-based Sybil Detection in Social and Information Systems has been accepted at the 2013 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM’13), which had 13% acceptance rate.
In this work, we propose a uniﬁed framework for systematic evaluation of Graph-based Sybil Detection (GSD) algorithms. We used this framework, in addition to a trace of a real-world Sybil activity in Facebook, to show that GSD algorithms should be designed to ﬁnd local community structures around known non-Sybil identities, while incrementally tracking changes in the graph as it evolves over time.* This introduces a paradigm shift from traditional graph statistics to dynamics.
We are currently investigating different ways to design an incremental GSD algorithm that scales to graphs consisting of hundreds of millions of nodes, while providing near real-time detection of Sybil identities with formal security guarantees. We are considering Apache Giraph for our system implementation. More awesome work is in-progress!
* SyPy is a Python implementation of the framework, and it is available here.