REVIEW
Data Rejuvenation: Exploiting Inactive Training Examples for Neural Machine Translation
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Data Rejuvenation: Exploiting Inactive Training Examples for Neural Machine Translation
read the original abstract
Large-scale training datasets lie at the core of the recent success of neural machine translation (NMT) models. However, the complex patterns and potential noises in the large-scale data make training NMT models difficult. In this work, we explore to identify the inactive training examples which contribute less to the model performance, and show that the existence of inactive examples depends on the data distribution. We further introduce data rejuvenation to improve the training of NMT models on large-scale datasets by exploiting inactive examples. The proposed framework consists of three phases. First, we train an identification model on the original training data, and use it to distinguish inactive examples and active examples by their sentence-level output probabilities. Then, we train a rejuvenation model on the active examples, which is used to re-label the inactive examples with forward-translation. Finally, the rejuvenated examples and the active examples are combined to train the final NMT model. Experimental results on WMT14 English-German and English-French datasets show that the proposed data rejuvenation consistently and significantly improves performance for several strong NMT models. Extensive analyses reveal that our approach stabilizes and accelerates the training process of NMT models, resulting in final models with better generalization capability.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.