Skip to last block for actual question, everything else is about what i see and dont understand.

Machine learning and artificial intelligence is very interesting to me, ive watched a few videos but i cant manage to wrap my mind around it.

I see a few people starting out with projects that appear to be an easy start, but i of course have no idea, were they make a self-driving car in GTA (crashes alot, but still) or teach the program to complete levels in a game (snake, mario, run forrest)

I watched a few videos on Jabrils youtube channel that seemed to make alot of sense, until one point..

How does the AI know when it hit a wall?
How does it know where the walls are?
How does it measure the distance? How does it know when it has respawned?

I find it really, really confusing.
Can anyone of you geniuses suggest me anything to get into this? Id prefer if the goal was to make an AI using machine learning, that can complete some basic game, like in Jabrils videos.

  • 0
    Im not familiar with ai/ml but i assume you can write software to gather that data by accessing program memory in the same way a CSGO hack reads memory to find out where enemy players are on the map. This can be extended to read the location of any element that the engine is referencing, you can extend this to read action triggers like death or interactions. On the other hand you get plugins that make the whole thing easier for things like game emulators
  • 0
    @simpleJack that could be one possibility, i didn't think of that. Thanks for the reply :)
Add Comment