Tracking and analysing security trends presented at AusCERT 2014

by Matt Jones,

Summary : This presentation will take a look at the latest innovations in performing data-mining of social network information that will detail the technical implementation of the Talkback core; it will include demos and a look at recent case-studies.
The presentation will show how we approached the challenge of identifying trending research - i.e. links to research - pdf's, blog posts, etc. and essentially keeping a thumb on the pulse of what the community is doing. We adopted a number of approaches to ensure accuracy and efficiency with our mining processes, and will talk through some of the technical (including some academic theories) - such as:
social network graphing
ranking by centrality
community detection
clustering users by profile/self-descriptions
applying a bag-of-words model
signature extraction
We will also run through a number of interesting real-world results and analysis of trending topics in InfoSec over the past year, such as badbios, big incidents, vulnerability research trends, and technical security innovation. This allows us to analyse not only the timeline of these events, but also shows how the security community comes together to contribute and some interesting view-points that data gathering and data-mining can provide (demographic trends, etc.).
During the presentation we will demonstrate using the new version of Talkback (there is a free version and a paid version, but the features that will be discussed and showed will primarily be in the core engine, which both versions use) and intend to genuinely talk about the interesting insights captured and thought process when improving algorithms and applying new approaches to break down the data.