The MPG data science team have been secretly working on universal approximation theorems and the mathematical theory of artificial neural networks, to establish new dense algorithmically generated class functions for the games industry.

Our bespoke feedforward architecture uses an approximation of two Euclidean spaces, with respect to compact convergence topology, to generate a ‘random’ neural network as a universal approximator for the success of a games idea.

So far, our tests have produced a variety of results, all of which far exceed the predictive capabilities of convolutional neural quantifiers. This wide variety of interesting results can give games developers a greater than zero chance when predicting success from even the simplest formulation of a games concept.

What’s more, and in a truly ground-breaking breakthrough, the MLG AI only needs a hand-held console or high-end mobile phone to run over six quadrillion neurons per second.

Come and join the AI revolution, see the future, find out what will be fun and what won’t as soon as you’ve written it down!