You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
But good procedural generation should give rise to emergent interestingness: fundamentally surprising and engaging dynamics that weren't explicitly placed there by a human designer. Genetic algorithms will give you that. Endless forms most beautiful...
Remember that the goal of procgen is not uniqueness via randomness, but uniqueness via novelty, i.e. interestingness generation. Emergence.
I think a good simulation is key for interestingness(where "good" is not necessarily powerful) The latest Zelda game for example has some great emergent gameplay born from the interaction of a very simple physics system(for a modern game) and an even simpler "chemistry" system.
The opposite of the original post is what happened with No Man's Sky. It's nearly all ProcGen, but it was so shallow in gameplay, the uniqueness wore off very quickly.
This is still considered an open problem in ALife. Whether you get continuing open-ended novelty or not depends on aspects that are still hand-made (e.g. laws of physics, fitness function, resource context). How to make systems that have infinite emergence potential is unclear.
Animations like @RickNMorty have continuous changing scenes into different dimensions and worlds. Thats keeps it interesting.
A #JumpNN should be able to jump into different interesting scenes/universes using holes.
As a former game-dev, there are some on-going researches on emergence. "Fun" is generated when a player encounters predicted uncertainty, so if we can make/train "perception model" of each player, fitness(novelty/interestingness) could be measured(predicted) with those models.
Actually, those topics were my last project in the game industry (for next-gen massively multiplayer game)
The complexity lies in how to measure "interestingness", that requires a general understanding of the world. Something interesting generally refers to some knowledge outside the " game universe".
Is it possible to build an ASIC for gen algorithms? What kind of math do they run on.. I'm completely clueless about them besides the fact that they're 'greedy' optimisers as compared to RL
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
