Standard 9 stops to get here · leads to 1

Contrastive Learning

A self-supervised learning approach that learns representations by contrasting similar and dissimilar examples.

Your route here

9 stops · basics first
  1. Machine Learning ✓ understood

    Building systems that learn patterns from data instead of following hand-written rules, getting better at a task as they see more examples.

  2. Unsupervised Learning ✓ understood

    Learning from unlabeled data to discover hidden patterns, structures, or relationships without explicit target outputs.

  3. Self-Supervised Learning ✓ understood

    Learning representations from unlabeled data by creating supervised tasks from the data itself (masked prediction, contrastive learning).

  4. Dataset ✓ understood

    A collection of data examples used for training, validating, or testing machine learning models.

  5. Feature ✓ understood

    A single measurable property of an example, such as a house's floor area or how many links an email contains, used as an input to a model.

  6. Neural Network ✓ understood

    A computational model inspired by biological neural networks, consisting of interconnected nodes (neurons) organized in layers that process information through weighted connections.

  7. Deep Learning ✓ understood

    A subset of machine learning that uses neural networks with multiple layers (deep neural networks) to learn hierarchical representations of data.

  8. Representation Learning ✓ understood

    Learning useful features or representations of data automatically, rather than hand-crafting them.

  9. Embedding ✓ understood

    A list of numbers (a vector) that represents a word, sentence, image or other item, learned so that similar items end up close together.

  10. Contrastive Learning · you are here ✓ understood

Where it sits

Before this

Self-Supervised Learning Embedding
Contrastive Learning

Leads to

CLIP

Explore nearby

Training Contrastive Loss A loss function for learning similarity metrics, bringing similar pairs together and separating dissimilar ones. Training Triplet Loss A loss for learning embeddings by pulling similar examples together and pushing dissimilar ones apart. Vision & Multimodal CLIP Contrastive Language-Image Pre-training - a model jointly trained on images and text, enabling zero-shot image classification. Neural Networks Representation Learning Learning useful features or representations of data automatically, rather than hand-crafting them. Training Data Augmentation Creating variations of training data through transformations (rotation, cropping, noise) to improve model generalization.

In the research

All papers →

2 papers that build on Contrastive Learning .