Pre-training a BERT with Curriculum Learning by Increasing Block-Size of Input Text
Koichi Nagatsuka, Clifford Broni-Bediako, Masayasu Atsumi
Abstract
Recently, pre-trained language representation models such as BERT and RoBERTa have achieved significant results in a wide range of natural language processing (NLP) tasks, however, it requires extremely high computational cost. Curriculum learning (CL) is one of the potential solutions to alleviate this problem. CL is a training strategy where training samples are given to models in a meaningful order instead of random sampling. In this work, we propose a new CL method which gradually increases the block-size of input text for training the self-attention mechanism of BERT and its variants using the maximum available batch-size. Experiments in low-resource settings show that our approach outperforms the baseline in terms of convergence speed and final performance on down-stream tasks.