Title:
Leveraging Social Networks to Fight Spam


Abstract:
Email communication contains an image of the human social network.
Emails sent to multiple recipients have some information about the
structure of that network.  In this talk I will describe a method of
spam filtering based on a network structure embedded in email messages.
This approach leverages the fact that email networks are formed by
social processes, which have certain signatures which are visible in the
graph structure.  I will outline how this technique may be used to
bootstrap other text classification techniques.