The Nous Institute’s foray into the world of artificial neural networks differs from the standard goal of exceeding human capacity through artificial intelligence.
Instead, our models are built and trained with one goal in mind; we aim to provide platforms that trigger novel human thought within the minds of users.
We believe that civilizational development occurs through human flourishing, and that growing dependence on AI is currently headed towards a reduction in general innovative capacity. Our family of machine learning models aims to transform machine learning into a tool that cultivates creativity in human beings, working in symbiosis with human ideas, rather than existing as a more efficient replacement.
FELIX, our set of GAN models, is trained on the works of fine artists and designers that collaborate with us. Our goal with FELIX is to catalog characteristics from the models' outputs that retain fidelity to the individual artist's aesthetic. Models are often trained on various artists' works in order to create fluid collaborations. Through our efforts with FELIX, the visual humanities are transformed into something other than human; they become canvases imbued with this "otherness" for artists to use as a foundation for their work.
DELPHI explores the human ability to synthesize knowledge through semiotic form. A set of finetuned GPT-J-6B model trained on custom datasets, DELPHI provides semiotics that are canonically resonant with the input material, allowing readers to draw meaning using that information in order to form new philosophical conclusions. DELPHI presents with significant consistency, and is gradually evolving in both scope and translatability. In theme with its namesake, DELPHI attempts prophecy, uttered under a digital trance and suspect to our hermeneutics.