Aiconomy

Self-Supervised Learning

A training paradigm where AI models learn from unlabeled data by creating their own supervisory signals, such as predicting masked words or future frames in video.

Self-supervised learning has been called the key to unlocking AI's potential by Yann LeCun. It eliminates the need for expensive manual labeling — critical given that less than 1% of the world's data is labeled. Both BERT (masked language modeling) and GPT (next-token prediction) are self-supervised approaches. The technique has enabled training on internet-scale datasets of trillions of tokens. Self-supervised pre-training followed by supervised fine-tuning has become the dominant paradigm for building state-of-the-art AI systems.

Explore the Data

AI Economy Pulse

Every Friday: the 3 AI data points that actually matter this week. Free, forever.

Built on data from Stanford HAI, IEA, OECD & IMF

Latest: “AI Investment Hits $42B in Q1 2026 — Here's Where It Went”

No spam, ever. Unsubscribe anytime.