Hacker News new | ask | show | jobs
by jrockway 2790 days ago
I am guessing that they look for cheating based on actions in addition to scanning the memory looking for known cheats. Aimbots are pretty recognizable to experienced players, and no doubt some machine learning is going on to look for those patterns. This lets them stay ahead of the curve on exploits... if it looks like an aimbot, it doesn't matter that it's some new binary they haven't yet heard of.
1 comments

For some reason it never occurred to me that they'd use ML for detecting suspicious behaviour, until I read your comment.

I've never considered this but it makes perfect sense, thanks for that!

https://www.youtube.com/watch?v=ObhK8lUfIlc

Watch this! It's about using ML to catch cheaters.

I searched youtube to show you this video; I used the search term "cs go machine learning". The first video was actually this: https://www.youtube.com/watch?v=w88RIcTuGZQ

Looks like it's using ML to make better aimbots--lol!