Deep learning, a/k/a deep neural network, is an aspect of Artificial Intelligence (AI) that emulates the learning process that humans use to gain certain types of knowledge.  Traditional machine learning algorithms are linear.  Deep learning algorithms, however,  are stacked in layers of increasing complexity and abstraction like layers on an onion.  For example, a toddler’s first word may be cat.  The child points to objects and says the word cat.  That parents says, Yes, that is a cat, or No, that is not a cat.  As the toddler continues to point to objects, s/he becomes more aware of the features that all cats possess.  The toddler learns what a cat is and is not.  With experience, the child clarifies a complex abstraction by building a hierarchy in which each level of abstraction is created with knowledge gained from the preceding layer of the hierarchy.  Deep learning in AI creates layers of meaning, each with increasing clarity and specificity.

As a trainer, we do the same thing with a learner.  The learner may learn what an arc welding rod is, but soon learns that this new one is an E-6013 arc welding rod.

The forthcoming lesson on Computer Modeling builds on such concepts.