ClassroomThe classroom
Back to the lesson libraryMECHANISM · 13 MIN
M12.6 CONNECT THE MECHANISM

Learn visual features before choosing the final task

You have 200 labeled receipt photos and millions of unlabeled ones. Learn how models teach themselves from unlabeled pictures, by matching cropped views or filling in hidden patches.

LESSON OVERVIEW13 min lesson

Lesson overview

You have 200 labeled receipt photos and millions of unlabeled ones. Learn how models teach themselves from unlabeled pictures, by matching cropped views or filling in hidden patches.

What you’ll explore

  • Self-supervised visual objectives learn from transformations, missing patches, or paired views; transfer evaluation determines whether the representation supports the intended visual task.
GO TO THE SOURCE

Original explanations, connected to the research.

A Simple Framework for Contrastive Learning of Visual Representations (SimCLR; Chen, Kornblith, Norouzi & Hinton, 2020)Masked Autoencoders Are Scalable Vision Learners (He et al., 2021)Learning Transferable Visual Models From Natural Language Supervision (CLIP; Radford et al., 2021)Dive into Deep Learning — fine-tuning
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.