This course introduces artificial intelligence (AI) programming tools inspired by our understanding of the human brain. The course includes four programming assignments in Python covering the four units of the brain as proposed by the Massachusetts Institute of Technology Center for Brain, Minds and Machines (CBMM): sensor stream, brain operating system, cognitive core, and symbolic compositional models. Collectively the four assignments introduce a set of tools and computer science concepts, with a focus on deep learning, spanning a basic skill set to program complete models able to perform AI tasks. Part of the assignments include comparing the deep learning tools implemented with other AI tools not based on neural networks. The focus of the assignments is to build models reproducing as closely as possible the complex cognitive tasks humans do naturally. Human intelligence can be characterized in a variety of ways and as part of the course, we review how various computer engineering applications may benefit from these different advances in modeling human intelligence. We discuss various integrative approaches aiming at combining experimental techniques in neuroscience and cognitive science, with computational modeling in order to elucidate the architecture of intelligence. The course provides background to understand some of the current limitations in our progress towards a general artificial intelligence machine.
Harvard Division of Continuing Education