On The Power of Curriculum Learning in Training Deep Networks

TL;DR

This paper investigates how curriculum learning, which involves non-uniform data sampling based on difficulty, can accelerate training and improve the performance of deep neural networks, especially CNNs for image recognition.

Contribution

It introduces methods for sorting training examples by difficulty and guides sampling with pacing functions, supported by empirical and theoretical analysis.

Findings

Curriculum learning increases training speed.

It improves final test performance.

Theoretical analysis shows it modifies the optimization landscape.

Abstract

Training neural networks is traditionally done by providing a sequence of random mini-batches sampled uniformly from the entire training data. In this work, we analyze the effect of curriculum learning, which involves the non-uniform sampling of mini-batches, on the training of deep networks, and specifically CNNs trained for image recognition. To employ curriculum learning, the training algorithm must resolve 2 problems: (i) sort the training examples by difficulty; (ii) compute a series of mini-batches that exhibit an increasing level of difficulty. We address challenge (i) using two methods: transfer learning from some competitive ``teacher" network, and bootstrapping. In our empirical evaluation, both methods show similar benefits in terms of increased learning speed and improved final performance on test data. We address challenge (ii) by investigating different pacing functions to guide the sampling. The empirical investigation includes a variety of network architectures, using images from CIFAR-10, CIFAR-100 and subsets of ImageNet. We conclude with a novel theoretical analysis of curriculum learning, where we show how it effectively modifies the optimization landscape. We then define the concept of an ideal curriculum, and show that under mild conditions it does not change the corresponding global minimum of the optimization function.