Discriminative vs. Generative

From
Revision as of 20:14, 5 January 2019 by BPeat (talk | contribs)
Jump to: navigation, search

YouTube search... ...Google search


Model Contrasts...

  • Discriminative
    • learn the (hard or soft) boundary between classes
    • better performance with lots of example data
    • not designed to use unlabeled data


  • Generative
    • model the distribution of individual classes
    • decision boundry: where one model
    • natural use of unlabeled data