Can you provide high level details on the type of algorithms used to affect a persistent state? I ask because I think that is key to the execution at the UX lev

Can you provide high level details on the type of algorithms used to affect a persistent state? I ask because I think that is key to the execution at the UX level of apps in that platform. For example a semi dense slam layer on top of ARkit is worth it, but a corelocation/mapkit system is not because the precision is not high enough. (Although maybe if it's paired with a images recognition system to breach the gap in precision...)