CS231n Note

CS231n Note

Concepts

Concept	Description
Image Classification
Object Detection
Action Classification
Image Captioning
Semantic Segmentation
Perceptual Grouping

Cross Validation

Used to set hyper-parameters.
Used commonly in small data set, not so much in deep learning.

Test set may not be representative of data in the wild.

Rsolution:
Partition data randomly between train and test data.

K-nearest neighbor on images never used

• Very slow at test time.
• Distance metrics on pixels are not informative.
• Curse of dense dimensionality.