ClassroomThe classroom
Back to the lesson libraryMECHANISM · 10 MIN
M03.7 CONNECT THE MECHANISM

Create more examples without inventing evidence

Flip a photo of a cat and it's still a cat. Flip a "turn left" sign and it's a different sign. Learn when made-up examples help a model, and when they quietly teach it lies.

LESSON OVERVIEW10 min lesson

Lesson overview

Flip a photo of a cat and it's still a cat. Flip a "turn left" sign and it's a different sign. Learn when made-up examples help a model, and when they quietly teach it lies.

What you’ll explore

  • Augmentation and synthetic data alter the training distribution; their usefulness depends on label validity, diversity, provenance, and evaluation on independent target data.
GO TO THE SOURCE

Original explanations, connected to the research.

Dive into Deep Learning — image augmentationDatasheets for DatasetsAI models collapse when trained on recursively generated data (Shumailov et al., 2024)
Suggest a correction

A precise note can make an explanation better.

Choose the scene and describe what needs attention. Download a feedback file to share through a channel you already use. This page does not send feedback or connect you with a reviewer.

The file includes this note, the scene title, and lesson metadata. Your saved progress and quiz responses are excluded. Download before leaving or reloading to keep your note.