It seems no one is using these lists to announce new papers. Maybe it's good that we don't, because it's rude to spam mailing lists with paper announcements? Or maybe people appreciate it? Or maybe the truth is somewhere in between, that a paper announcement now and then is fine?
Anyhow, we have a new paper.
We are not generating levels, we are generating level generators (therefore "metageneration"). Through evolution with LLMs in the loop. And we abstract functions for reuse, that helps later level generators be better and smaller. Methods like assuring that there is exactly one player, adding balanced platforms, etc. This is a bit like automatically defined functions from good old Genetic Programming, but taken into the LLM era.
--