CAMeMBERT: Cascading Assistant-Mediated Multilingual BERT

TL;DR

CAMeMBERT introduces a knowledge distillation approach to create a more efficient multilingual BERT model, reducing resource requirements while maintaining acceptable accuracy levels for NLP tasks.

Contribution

The paper presents CAMeMBERT, a novel cascading distillation method that enhances multilingual BERT's efficiency with minimal accuracy loss.

Findings

Achieves around 60.1% accuracy on NLP tasks.

Reduces time and space complexity compared to original mBERT.

Uses a cascading distillation process with teacher assistant networks.

Abstract

Large language models having hundreds of millions, and even billions, of parameters have performed extremely well on a variety of natural language processing (NLP) tasks. Their widespread use and adoption, however, is hindered by the lack of availability and portability of sufficiently large computational resources. This paper proposes a knowledge distillation (KD) technique building on the work of LightMBERT, a student model of multilingual BERT (mBERT). By repeatedly distilling mBERT through increasingly compressed toplayer distilled teacher assistant networks, CAMeMBERT aims to improve upon the time and space complexities of mBERT while keeping loss of accuracy beneath an acceptable threshold. At present, CAMeMBERT has an average accuracy of around 60.1%, which is subject to change after future improvements to the hyperparameters used in fine-tuning.