Introduction to Machine Learning (Fall 2020) (M-I-T)

Back-propagation through time – backwards pass (M-I-T)

Back-propagation through time – forward pass (M-I-T)

Back-propagation through time – weight updates (M-I-T)

Building a tree – greedy algorithm (M-I-T)

Demo example – RNNs (M-I-T)

Building a tree – minimum error splits (M-I-T)

Evaluating hypotheses – training set error (M-I-T)

Building a tree – pruning (M-I-T)

Evaluating learning algorithms – validation (M-I-T)

Classification trees (M-I-T)

Evaluating predictions – loss functions (M-I-T)

Classification trees – impurity measures – entropy (M-I-T)

Example of perceptron algorithm with polynomial basis transformations (M-I-T)

Classification trees – impurity measures – gini index (M-I-T)

Feature representation – polynomial basis (M-I-T)

CNNs – a specific illustrative example filter (M-I-T)

Feature representation – transforming through-origin to not-through-origin (M-I-T)

CNNs – backprop and gradient descent (M-I-T)

Feature representation strategies for dealing with varied data (M-I-T)

CNNs – convolutional neural network layers (M-I-T)

Gradient descent optimization – algorithm in multiple dimensions (M-I-T)

CNNs – convolutional neural networks – intro (M-I-T)

CNNs – max pooling (M-I-T)

Gradient descent optimization – algorithm in one dimension (M-I-T)

Gradient descent optimization – local optima (M-I-T)