Knowledge Distillation for Quality Estimation

Amit Gajbhiye, Marina Fomicheva, Fernando Alva-Manchego, Frédéric Blain, Abiola Obamuyide, Nikolaos Aletras, Lucia Specia

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

16 Downloads (Pure)

Abstract

Quality Estimation (QE) is the task of automatically predicting Machine Translation quality in the absence of reference translations, making it applicable in real-time settings, such as translating online social media conversations. Recent success in QE stems from the use of multilingual pre-trained representations, where very large models lead to impressive results. However, the inference time, disk and memory requirements of such models do not allow for wide usage in the real world. Models trained on distilled pre-trained representations remain prohibitively large for many usage scenarios. We instead propose to directly transfer knowledge from a strong QE teacher model to a much smaller model with a different, shallower architecture. We show that this approach, in combination with data augmentation, leads to light-weight QE models that perform competitively with distilled pre-trained representations with 8x fewer parameters.
Original languageEnglish
Title of host publicationACL Findings 2021
Publication statusPublished - 1 Jul 2021
Externally publishedYes

Keywords

  • cs.CL

Fingerprint

Dive into the research topics of 'Knowledge Distillation for Quality Estimation'. Together they form a unique fingerprint.

Cite this