Thanks! You can teach a computer to perform a very small task, if you give it a huge amount of data (designed with that task in mind) and find a way to tell it when its wrong. That last part is the hardest. That's the crux of the problem. How it works is still a bit of dark magic.
You are viewing a single comment's thread from:
i seen some of how that works as i watched some videos from Comma ai where people were fixing masks on photos that ai uses for recognizing cars, road, grass...
it is interesting.