May 2018
Intermediate to advanced
576 pages
14h 42m
English
As explained in the previous section, semi-supervised learning is not guaranteed to improve a supervised model. A wrong choice could lead to a dramatic worsening in performance; however, it's possible to state some fundamental assumptions which are required for semi-supervised learning to work properly. They are not always mathematically proven theorems, but rather empirical observations that justify the choice of an approach otherwise completely arbitrary.
Read now
Unlock full access