EasyDeepLearn

How does Mixup work?

medium

Answer

  • Sample two examples (xi,  yi)(x_{i}, \;y_{i}), (xj,  yj)(x_{j}, \;y_{j}) and a mixing coefficient λ ~ Beta(α, α) with α ~ 0.2.
  • Train on the interpolation: xnew  =  λ⋅xi  +  (1−λ)⋅xjx_{\mathrm{new}}\; = \;{\lambda} \cdot x_{i}\; + \;(1 - {\lambda}) \cdot x_{j}, ynew  =  λ⋅yi  +  (1−λ)⋅yjy_{\mathrm{new}}\; = \;{\lambda} \cdot y_{i}\; + \;(1 - {\lambda}) \cdot y_{j}.
  • Forces linearity between examples — smoother decision boundaries, better calibration, robustness to adversarial noise.
  • Widely used in vision.
  • Doesn't work as well on text without variants.
Check yourself — multiple choice
  • Mixup concatenates two images
  • Linearly interpolate inputs and labels with λ ~ Beta — smoother decision boundaries and better calibration
  • Only augments images, not labels
  • Same as CutMix

Mixup: linearly blend both inputs and labels → smoother, better-calibrated decisions.

#augmentation#mixup#regularization

Practise Deep Learning

214 interview questions in this topic.

Read the lessonFlashcardsTake the quiz

Related questions