Descending Into ML Flashcards
1
Q
What is “Training” a model?
A
learning (determining) good values for all the weights and the bias from labeled examples.
2
Q
What is “empirical risk minimization” ?
A
attempting to find a model that minimizes loss
3
Q
What is “squared loss” ?
A
= the square of the difference between the label and the prediction
= (observation - prediction(x))2
= (y - y’)2
4
Q
What is Mean square error (MSE) ?
A
the average squared loss per example over the whole dataset.
5
Q
How to calculate MSE ?
A
sum up all the squared losses for individual examples and then divide by the number of examples:
