While true Artificial General Intelligence is still far away, in the short term we can augment machine learning with ‘general’ capabilities and features. Unlike most AI/ML, general intelligence should not require tailoring to specific problems. Instead, it learns continuously from interactions with a world, developing an internal representation by exploring and exploiting what it already knows. Algorithms with general intelligence features are applicable to a wider range of problems and tasks.
Sep 3rd 2018: New RFRs published on Bicameral Machine Learning models and self organising architectures.
Apr 18th 2018: Released ArXiv paper “Sparse Unsupervised Capsules Generalize Better”