FRAME Project

FRAME (ReFRAMing MachinE Learning via Ordinal Tensors and Deep Quality Diversity) is a Maltese project funded by MCST and is a "sister" project of TAMED.

FRAME will exploit the tensor-based Machine Learning models developed during TAMED and enhance them with Ordinal Machine Learning algorithms and Quality Diversity techniques. FRAME envisages coupled tensor-based models with divergent search for yielding both diverse (novel, surprising) and high quality outcomes for both problem solving and generative tasks.

Digital Games will be used for evaluating the outcomes of FRAME since: 1) it is empirically evident by core publications in leading AI conferences that games provide a unique test-bed for testing AI to its limits across dissimilar tasks, 2) games provide the sole domain where problem solving in highly deceptive environments (game playing) meets artefact generation (game design) and computational art, and 3) both games and AI define core strategic areas of development for Malta.

Read more