# What is Self-Supervised Learning? Definition & Technical Mechanics

> **Source:** [SPIDITS AI AI Glossary](https://spidits.com/ai-glossary)  
> **Category:** Model Training  
> **Canonical Citation:** [https://spidits.com/ai-glossary/self-supervised-learning](https://spidits.com/ai-glossary/self-supervised-learning)

## Definition
Self-Supervised Learning is a training paradigm where the model generates its own labels directly from the input data (e.g. masking words and predicting them), allowing training on massive unlabeled datasets without human labeling.

## Architectural Mechanics & Deep Dive
Self-supervised learning is a training paradigm where a model generates its own training labels from unlabeled raw data. Common tasks include Masked Language Modeling (predicting hidden words) and contrastive learning (matching modified views of an image). Self-supervised learning is the key driver of generative AI, allowing models to pre-train on massive raw datasets.

## Industry Applications & Real-World Use Cases
Foundation model pre-training, BERT training cycles, and visual representations

## Frequently Asked Questions

### How does self-supervised learning compare to supervised learning?
Supervised learning requires human annotators to label images or text. Self-supervised learning automates labeling by hiding parts of the input data and forcing the model to guess them.

### Is LLM pre-training self-supervised?
Yes, because the next-word prediction objective automatically uses the subsequent words in the dataset as the labels.

## Related Terminology
- [supervised-learning](https://spidits.com/ai-glossary/supervised-learning)
- [unsupervised-learning](https://spidits.com/ai-glossary/unsupervised-learning)
- [llm](https://spidits.com/ai-glossary/llm)

---
*From the SPIDITS AI Knowledge Base (336+ terms indexed). Explore full technical definitions: [https://spidits.com/ai-glossary](https://spidits.com/ai-glossary)*