Deep Learning

DL concepts

DL is a subset we’ve discussed at length throughout this course, and it relies on neural networks to arrive at insights and predictions. We’ve likely seen the term opaque-box models thrown around often throughout the course, and this is our reminder that this likely means DL models are involved because hidden layers are opaque structures that the humans that program DL neural networks can’t see into. The reason for this is that within each hidden layer, there are nodes that make computational decisions, and these decisions are happening essentially all at once, so it’s not a matter of not wanting to offer transparency into the models—it’s a matter of really not being able to in the first place.