Zubnet AILearnWiki › Curriculum Learning
Training

Curriculum Learning

A training strategy that presents examples in a meaningful order — typically from easy to hard — rather than randomly. Like teaching a student arithmetic before calculus, curriculum learning gives the model foundational patterns first and builds complexity gradually. This can lead to faster convergence and sometimes better final performance.

Why it matters

Curriculum learning is an underappreciated technique that can improve training efficiency without changing the model or data. LLM pre-training increasingly uses data scheduling — showing cleaner, higher-quality data in the final training stages — which is a form of curriculum learning. The order you present data matters, not just the data itself.

Deep Dive

The intuition: random training order means the model encounters hard examples before it has the foundation to learn from them, wasting gradient updates. By starting with easy, clear examples, the model builds basic representations that make harder examples learnable. Bengio et al. (2009) formalized this, showing that curriculum training converges faster and sometimes reaches better solutions than random ordering.

In LLM Pre-training

Modern LLM training uses a form of curriculum learning through data mixing schedules. Early training might use a broad mix of web text. Later stages increase the proportion of high-quality data (curated text, code, math, reasoning). Some training recipes "anneal" on premium data in the final phase — the model learns broad patterns first, then refines on carefully selected examples. This data scheduling is one of the most closely guarded secrets of frontier labs.

Self-Paced Learning

Self-paced learning is a variant where the model itself decides what's easy or hard, based on its current loss on each example. Examples with low loss (already learned) are deprioritized, while examples with moderate loss (challenging but learnable) are emphasized. Examples with very high loss (noise, mislabeled data) are also deprioritized as potential outliers. This adaptive curriculum adjusts automatically as the model improves.

Related Concepts

← All Terms
← Cross-Attention Cursor →