Hey OldAccount29, I didn’t see your replies until now. Somehow the thread subscription thingee didn’t send me an email like it usually does!
I can’t promise I’ll read every single link in that list I’m afraid! This week has been especially hard on me with 16+ hour days trying to put together the first full-length gameplay video for DoN. It has so many features I’ve never shown before but it also revealed so many critical technical and performance issues that I’ve been racing non-stop around the clock to plug, fix and re-test, a very stressful time indeed.
I’ll try to check out the videos to see what the competition in this genre is like though 
If someone had given me these links 1.5 years ago, I could have taken the game in a different direction, but it’s probably too late now. Machine learning sounds interesting for sure, but I’m already sitting atop a frightening pile of unfinished features, bugs, and performance issues so I simply can’t afford to take on new features!
Of course - you have a fresh start ahead of you so make the most of all this research you’ve unearthed! In my experience it is best to define functional requirements clearly before the project begins and stick to it (while being nimble where necessary) instead of making giant functional leaps from time to time like I have done; I’ve thrown away much code and assets from time to time it makes my head spin when I think about it lol.
BTW I appreciate your unbridled enthusiasm, this is a very interesting genre for sure and there’s so many different approaches and sub-genres within it, some day when my game is complete I just want to sit back, relax and play all these lovely games of the natural world one by one 