Virus Signatures and Genetic Algorithms

Question

I would like to know how one achieves the following signature. I have read online that (al least in the past) researchers will take the "suspected" file the binary code, convert it to assembly, examine it, pick sections of code that appear to be unusual, and identifying the corresponding bytes in the machine code.

But then how is the bellow virus string signature achieved?

MIRC.Julie=6463632073656e6420246e69636b20433a5c57696e646f77735c4a756c696531362c4a50472e636f6d0a0d6e31333d207d0a0d6e31343d200a0d6e31353d206374637020313a70696e673a2f6463632073656e6420246e69636b20433a5c57696e646f77735c4a756c696531362c4a50

Also, (although this might sound completely crazy) that string above must mean something, i can only guess a sequence of actions, actual code, etc. So if it was once "translated" in this form (virus signature) from assembly, is it possible to convert it back?

---

Just in case you might wonder why am asking what even I think is a weird question. This is why... I am preparing my BSc final year computer science project, and at this point I am wondering whether it would be possible to maybe generate/estimate/evaluate/predict virus signatures by using GA's (Genetic Algorithms). Maybe that will help make my question a bit easier to understand, I hope.

Thanks!

Answer 1

The virus signature shown is probably dependent on the scanner that generated it. I find it extremely easy to believe that all virus scanners create their signatures in different ways. Without a source, there's no way to explain how it was developed, and even with a source I doubt this is something that AV companies will reveal, since it allows virus developers the opportunity to avoid detection.

"Generate/estimate/evaluate/predict" are four different problems and not all of them are best done with a GA. You need to select your problem before selecting an algorithm.

Answer 2

You cannot revert it back because normally virus signatures are encrypted. The way they are obtained is by extracting the binary mallicious code from an executable and then converting it to hexadecimal representation. Hopefully helps