Immune Networking - Artificial Immune Systems Approach

Routing Misbehavior Detection

In [C1] we apply a basic AIS algorithm - negative selection, and build a system (implemented in a simulator) which detects misbehavior of mobile nodes with respect to a routing protocol for building collaboratively an ad hoc network. We add a method to minimize delay of the detection decision under the maximum-allowed expected false-detection constraint.

In [J1] we improve [C1] by adding the AIS-memory component to the system, and show how this speeds up the secondary response (faster detection of repeated misbehavior).

In [C2,J2] we improve [C1,J1] by introducing a new concept that, along with the use of the negative feedback (communication problems) from the protected system, enables the negative selection to become dynamic – the system learns not-previously-defined normal behavior of the protected system and adapts to its changes. 

Go Up

Email Spam Filtering

We develop an artificial immune system for collaborative email spam filtering. In [C4, R1] we outline its design and initial evaluation. We filed a patent for our solution [P1]. Added to an email server, the system processes locally incoming emails and optionally collaborates to other such systems to better capture spam based on its bulkiness. It takes into account (using AIS algorithms) communication profiles of the users and explicit (delete as spam button) or implicit (users actions on received emails) feedback from the users. It is designed to be able to detect new spam bulk (evaluated, [C4, R1]) and repeated spam experienced by a specific user (to be evaluated), and to be resistant to heavy spam obfuscations (partially evaluated, [C4, R1]). We also developed a system for easy and automated evaluation of email spam filters [C3, S1]. We currently work on a novel component for securing communications that we think should improve our antispam system, but also be usable for some other communication problems  [not yet published].

Go Up

Publications

*** CONFERENCE PAPERS:

[C4] Artificial Immune System For Collaborative Spam Filtering. Slavisa Sarafijanovic and Jean-Yves Le Boudec. In Proceedings of NICSO 2007, The Second Workshop on Nature Inspired Cooperative Strategies for Optimization, Acireale, Italy, November 8-10, 2007.  [to appear] [Technical report version - PDF]

[C3] AntispamLab – A Tool for Realistic Evaluation of Email Spam Filters. Slavisa Sarafijanovic, Luis Hernandez, Raphael Naefen, and Jean-Yves Le Boudec. In Proceedings of CEAS 2007, The Fourth Conference on Email and Antispam, Mountain View, California, USA, p. 121-127, August 2007. [PDF] [Software of the tool]

[C1] An Artificial Immune System Approach to Misbehavior Detection in Mobile Ad-Hoc Networks. Jean-Yves Le Boudec and Slavisa Sarafijanovic. In Proceedings of Bio-ADIT 2004, The First International Workshop on Biologically Inspired Approaches to Advanced Information Technology, Lausanne, Switzerland, p. 96-111, January 2004. [PDF]

*** JOURNALS:

[J2] An Artificial Immune System for Misbehavior Detection in Mobile Ad-Hoc Networks with Virtual Thymus, Clustering, Danger Signal and Memory Detectors. Slavisa Sarafijanovic and Jean-Yves Le Boudec. In International Journal of Unconventional Computing, vol. 1, p. 221-254, 2005. [PDF]

*** PATENTS:

*** TECHNICAL REPORTS: