Neural Networks Intro: Perceptron Model - [2020-08-04]

The beginning of introducing neural networks. In future videos, we'll discuss implementation using Keras & TensorFlow, optimization techniques, and deep learning. In this video, we motivate the structure of neural networks by starting with the perceptron model to model logical operators. We then proceed to add more complexity (multiple layers) and introduce the terminology and use cases of a neural network. I mentioned for "homework" before the next Study Group on neural networks to watch this playlist from 3Blue1Brown: https://www.youtube.com/playlist?list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi 0:00 Overview structure of future Study Groups 1:27 Introducing neural networks 2:58 Perceptron model 5:20 Logic gates implementation 12:20 Multiple perceptrons: neural network 13:02 Hidden layer for latent features 16:56 Mathematical representation: Loss function 20:04 Relation to linear/polynomial regression? 22:43 TensorFlow playground: Visualizing learning process 27:40 Looking to next time 28:30 When deep learning is and isn't a solution 29:38 Analogy: Neural networks are children 30:30 Next time: Activation functions & hyperparameters 31:10 Homework for next time Notebook(s) can be found on https://github.com/MrGeislinger/flatiron-school-data-science-curriculum-resources