AI & MLCoreAIML-DATA-010
Labels: Annotation, Label Noise, and Weak Supervision
Where labels come from, how noisy labels bend what a model learns, and cheap alternatives to hand labeling.
- Difficulty
- 2 / 5
- Importance
- 5 / 5
- Study time
- 1–2 h
- Kind of work
- Mixed
Why it matters
The learner can vary a label-noise rate and predict how the learned boundary and test accuracy degrade, then name one weak-supervision source.