Curriculum learning

TLDR

Learning is more effective when examples are presented in a meaningful, progressively complex order, and curriculum learning is hypothesized to accelerate convergence and improve local minima quality in non‑convex training. The study formalizes curriculum learning as a machine‑learning training strategy. The authors investigate curriculum learning across different setups for deep deterministic and stochastic neural networks with non‑convex training criteria. Experiments demonstrate that curriculum learning yields significant generalization improvements, accelerating convergence and enhancing local minima quality.

Abstract

Humans and animals learn much better when the examples are not randomly presented but organized in a meaningful order which illustrates gradually more concepts, and gradually more complex ones. Here, we formalize such training strategies in the context of machine learning, and call them "curriculum learning". In the context of recent research studying the difficulty of training in the presence of non-convex training criteria (for deep deterministic and stochastic neural networks), we explore curriculum learning in various set-ups. The experiments show that significant improvements in generalization can be achieved. We hypothesize that curriculum learning has both an effect on the speed of convergence of the training process to a minimum and, in the case of non-convex criteria, on the quality of the local minima obtained: curriculum learning can be seen as a particular form of continuation method (a general strategy for global optimization of non-convex functions).