Week 5 (Gaussians) Flashcards
What are the 2 steps of the k-means algorithm
What is a gaussian mixture distribution (and how does it work mathematically)
What does the k means algorithm look like in practice
What is the responsibility function in a gaussian mixture distribution
What does soft clustering with gaussian mixtures look like in practice
How does MLE work for a gaussian mixture
What is the k-means distortion measure
Always goes down or converges
What is the premise behind the EM algorithm for gaussian mixtures
Does the E step use bases theorem
Yes
What does the EM algorithm look like in practice
What are the latent variables for the EM algorithm
The Z values reflect the assignments of responsibility (and are thus learned)
What do these values represent
What is the KL divergence
What are the key ideas behind the EM algorithm
What is the E step of the EM algorithm
Update the responsibilities
What is the M step of the EM algorithm
Compute the new parameters
What do responsibilities define in the EM algorithm
