Emergent Turing Machines and Operating Systems for Brain-Like Auto-Programming for General Purposes

In Artificial Intelligence (AI) there is a wide gap between the symbolic school and the connectionist school, but in both schools, different task-purposes require different learning methods. Different sensory modalities, such as video, sound, and text, also require different learning methods. In a larger scale, there is further a wide gap between brain scene and computer science — human brains automatically generate programs for general purposes but computers still cannot do so. Our Developmental Networks (DN) here are meant for bridging the gap in AI by further bridging the larger gap between brain science and computer science. The DN learning engine automatically emerges Turing Machine logic into its neural network. The AI Machine Learning (AIML) Contest 2016 is the first machine-learning contest that used task-nonspecific and modality-nonspecific learning engines. It used the DN engine for a variety of tasks and modalities. The organizers and contestants independently verified a prototype of DN for vision, audition, and natural languages acquisition and understanding (English and French co-acquisition). The Auto-programming Operating Systems (AOS) developed by GENISAMA LLC is meant to facilitate developers to train for many different applications using the same DN engine by following a standard for the body setting