Malware Detection in Executables Using Neural Networks

Re:Malware Detection in Executables Using Neural Networks by garganzol on Monday December 04, 2017 @01:10PM (#15846033)

As a desktop software vendor, I gathered a lot of perspective on this topic. AI anti-virus engines tend to mark everything as malware. Literally everything, even the simplest well-behaving "Hello World" apps. It looks like a pure manifestation of greed and laziness to me: smash a black box, pretend it does all the heavy lifting and relax taking all the orders you can take in this life.

More than that. Currently there are a lot of shady security suppliers who do nothing except imitating a good protection. They put fancy design into UI coupled with a mediocre 3rd party anti-virus engine, and presto, you have yet another source of false positives. They do not react to reports nor provide any meaningful support. Instead, they are pretty busy with marketing, buzz and sales. Yes, what could go wrong.

[ Parent | on HN ]

Re:Malware Detection in Executables Using Neural Networks by EdwardRaff on Monday December 04, 2017 @01:29PM (#15846212)

I'd be cautious about making broad statements regarding how AI based AVs tend to behave. It's a new space with a lot of competitors, of varying degrees of quality, with varying degrees of actual AI in use. Not all of them are even tackling the exact same problem, so comparisons can get murky fast.

Its also a space where data collection and ground truth are expensive, so people try to boot strap solutions in odd ways. Its tough.

As the paper author, I'm not going to name any companies. Especially since one was sharing data with us for this work, and I don't want to inject any of my biases into a particular name. But we've done some fairly extensive testing of the models we've built. Both on data they gave us and data we got ourselves, and our model don't just label everything as malware. It is not quite AV quality at the moment, but we also have limited training data and are really still exploring the different ways one can tackle this problem with AI.

Jon did a great job in this blog post, but if you want more details on why there is so much work to still be done, I'd recommend reading our paper! We tried to make the intro section accessible to people with no AV/malware background.

[ Parent | on HN ]

Re:Malware Detection in Executables Using Neural Networks by huhtenberg on Monday December 04, 2017 @04:03PM (#15847470)

> one was sharing data with us for this work

I sure hope it wasn't Tehtris, because their "AI detection" is nothing short of random.

[ Parent | on HN ]

Re:Malware Detection in Executables Using Neural Networks by lqdc13 on Monday December 04, 2017 @01:25PM (#15846172)

The reason simple hello world apps are marked as malware by AI-based engines is that small programs tend to be malware. There are probably close to no hello world programs in their dataset.

The problem is that to be a "good" program, yours has to do something other "good" programs do. And there aren't that many of those that use different software development paradigms (not different versions of the same one). So new different programs tend to be marked as malware.

This also brings up a very easy way to circumvent such AVs. Simply modify an existing goodware program and it will be marked as goodware. Add some obfuscation/polymorphism and it would be virtually impossible to detect such malware using static analysis/AI-based AVs.

[ Parent | on HN ]

Re:Malware Detection in Executables Using Neural Networks by cortesoft on Monday December 04, 2017 @04:49PM (#15847867)

Is there really malware that is as small as a hello world program?

[ Parent | on HN ]

Re:Malware Detection in Executables Using Neural Networks by lqdc13 on Monday December 04, 2017 @06:11PM (#15848453)

Yes, because they try to strip the binaries as much as possible so that the file size is smaller. If you compile a regular "hello_world.c" with gcc -O3, the size is 8 kb. You can def make a malware that is way smaller that does something simple like change some registry value to some uri.

Average size for malware is ~100kb-200kb btw. This is way smaller than almost any software besides some console games.

[ Parent | on HN ]

Re:Malware Detection in Executables Using Neural Networks by jstarfish on Monday December 04, 2017 @06:47PM (#15848626)

Lots of malicious Office macros and Javascript redirects come to mind. Less than 100K (minus the container) in most cases.

[ Parent | on HN ]