Compare three ways to reduce shortcuts: constrain the weights, mask units and vary the examples.
Proof of success: explain "Dropout, augmentation and regularization", complete "The Variant Lab" and name the limit of the analogy.
In a decoder team, every riddle is always solved by the same person. The group fails as soon as she's away. Lina imposes practice rounds where some randomly picked members stay silent: that's the idea of dropout. She also lightly rotates the patterns while keeping their label: that's a valid augmentation. Finally, she penalizes excessive settings. These tools reduce some overfitting, but only if their assumptions match the problem.
Preview — the rest of the lesson is for enrolled readers.
Your access is tied to your account, not to this link. Sign in with the same email you used in class: your course is waiting, no need to enter the code again.
Sign inNo account yet? Create oneThe first modules of the course are open to everyone. For the rest you have three options: buy this course once and for all, subscribe, or enter the code handed out in class.