Deep Learning Fundamentals

Deep Learning Fundamentals, course code ML0115EN, is a concept-only crash course produced by DeepLearning.TV and hosted on Cognitive Class. The page states an estimated effort of three hours, describes the course as free and self-paced with no prerequisites, and says it can be audited as many times as you wish; it does not state a level. Four modules of short concept videos run in order: deep learning concepts, covering what a neural network is, why deep learning matters, how to choose between networks, the vanishing gradient problem, restricted Boltzmann machines and deep belief networks; more concepts, including convolutional networks, recurrent nets, autoencoders, recursive neural tensor networks and use cases; platforms for deep learning; and deep learning software libraries.

The platform and library modules name tools such as H2O.ai, Dato GraphLab, Theano, Deeplearning4j, Torch and Caffe, so treat those two modules as historical context rather than a current tool survey. There is no coding and no maths beyond intuition; the aim is to show that the ideas are simpler than they sound before you pick a framework. The course is free of charge and the page says a certificate is offered on completion at no cost.