Papers by Everlyn Chimoto
COMET-QE and Active Learning for Low-Resource Machine Translation (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Using COMET-QE, we select sentences for low-resource neural machine translation. |
| Approach: | They propose a reference-free evaluation metric to select sentences for low-resource neural machine translation using Swahili, Kinyarwanda and Spanish. |
| Outcome: | The proposed method outperforms two variants of Round Trip Translation Likelihood and random sentence selection by up to 5 BLEU points on a 30k baseline. |
Critical Learning Periods: Leveraging Early Training Dynamics for Efficient Data Pruning (2024.findings-acl)
Copied to clipboard
| Challenge: | Neural Machine Translation models are extremely data-hungry and require a large dataset to maintain data quality. |
| Approach: | They propose a new data pruning technique that leverages early model training dynamics to identify the most relevant data points for model performance. |
| Outcome: | The proposed technique outperforms the benchmarks on indo-European languages while pruning up to 50% of training data. |