EasyDeepLearn

How does unsupervised learning help with noisy labels?

hard

Answer

  • (1) Cluster the data unsupervisedly; check consistency of labels within clusters — outliers likely mislabeled.
  • (2) Confidence learning (Northcutt et al., cleanlab): use self-consistency + probabilistic label pruning.
  • (3) Small-loss trick during training (Co-teaching): assume noisy samples have larger loss; ignore top-loss fraction.
  • (4) Semi-supervised: leverage unlabeled or auto-labeled data.
  • Foundation of modern label cleaning at scale.
Check yourself — multiple choice
  • Random
  • Cluster consistency check + cleanlab confidence learning + small-loss Co-teaching + semi-supervised leverage — modern label cleaning at scale
  • Same as supervised
  • Not real

Noisy labels: cluster consistency + cleanlab + Co-teaching small-loss + semi-sup.

#applications

Practise Unsupervised Learning

214 interview questions in this topic.

Related questions