The Anti-Spoofing Techniques within a TSF

The TSF allows for the use of dynamic recognition techniques (like our new techniques) since
there is no requirement of binding to real-world identities. Because participating users can
be recognised and not easily spoofed, a user can rely on his/her own observations to compute
its trustworthiness. However, the recognition is so low for non-participating users (who use
no added anti-spoofing protection) that it is not possible to compute an explicit trust value
in the senders based on past local interactions. Still, the TSF is useful due to its
collaboration feature, which is used to reduce uncertainty by making the knowledge of trusted
peers available to the anti-spam tool. For example, the collaboration features of the TSF may
also improve Bayesian filters – the TSF allowing the trustworthiness of collaborators to be
explicitly computed and evolve dynamically. So that if a misclassification due to the Bayesian
filter occurs, the incriminated email along with its correct classification (spam or non-spam)
may be pushed as a recommendation to other users. Based on the trust value of the recommender,
the receiver could add the embedded email to its local corpus of spam or anti-spam email
according to the embedded correct classification. Then, the Bayesian filter may be retrained in
order to be improved.

Concerning the implementations details of ER, if we take the example of the email system where
simple text email addresses are used for recognition, the ER process is mapped to:

  1. a new email is received;
  2. the text email address is compared to already stored email addresses;
  3. if this is a new email address, this one is optionally stored for convenience if replies are sent to this email address;
  4. the email is delivered in the Inbox folder of the user’s email client.

In our system, it is changed to the steps described in Fig. 1

Fig. 1: ER/TSF Global View



, , , , ,

  1. Leave a comment

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: