Interview

Page 1

T-SNE - t-SNE differs from PCA by preserving only small pairwise distances or local similarities whereas PCA is concerned with preserving large pairwise distances to maximize variance

PCA - PCA is a linear dimension reduction technique that seeks to maximize variance and preserves large pairwise distances

Page 2

SVM : Linear model with hinge loss

Read hinge loss - http://cs231n.github.io/linear-classify/

one ambiguity is if W gives 0 loss then 2W will also give

can be removed by Regularisation

l2 vs l1

l2 - more diffuse and smaller W

Note that biases do not have the same effect since, unlike the weights, they do not control the strength of influence of an input dimension. Therefore, it is common to only regularize the weights WW but not the biases bb.

Page 3