SIGBOVIK (ACH Special Interest Group on Harry Qatar Bovik) is an annual journal of parody machine learning papers whose 2019 proceedings came out today. Selectively, the parody is of very high quality and effort – it is worth replacing whatever you usually read on the toilet.
You can find the full version here, and some highlights are quoted bellow:
For all the PERL lovers:
In this paper, we aim to answer a long-standing open problem in the programming languages community: is it possible to smear paint on the wall without creating valid Perl?
Naturally, all implementations are solid (y):
We are confident in the underlying implementation because it relies on machine learning and neural networks, which are sufficiently sophisticated to ensure complete accuracy.
And, of course, quantified knowledge
Theorem 1 (Pictogram–kiloword equivalence theorem). A picture is worth a thousand words.
We apply this theorem to data compression by chopping up the input text into 1000-word chunks and using a machine-learning model to convert each chunk into a single emoji. […] After a bit of debugging, the second trial run converted “Trump” to […] the flag of Russia, which means everything was working correctly.