Current state-of-the-art NLP systems use large neural networks that require lots of computational resources for training. Inspired by human knowledge acquisition, researchers have proposed curriculum learning, - sequencing of tasks (task-based curricula) or ordering and sampling of the datasets (data-based curricula) that facilitate training. This work investigates the benefits of data-based curriculum learning for large modern language models such as BERT and T5. We experiment with various curricula based on a range of complexity measures and different sampling strategies. Extensive experiments on different NLP tasks show that curricula based on various complexity measures rarely has any benefits while random sampling performs either as well or better than curricula.
@article{arxiv.2112.06510,
title = {Do Data-based Curricula Work?},
author = {Maxim K. Surkov and Vladislav D. Mosin and Ivan P. Yamshchikov},
journal= {arXiv preprint arXiv:2112.06510},
year = {2024}
}