For the 3 directions (ende, zhen, heen) covered in WMT2023, the model is pre-trained on top of XLMR-L using synthetic data generated by DCSQE.
Jackie Lai
DreamW1ngs
AI & ML interests
synthetic data, LLMs inference, and multilingual LLMs.
Recent Activity
authored
a paper
6 days ago
Alleviating Distribution Shift in Synthetic Data for Machine Translation
Quality Estimation
authored
a paper
6 days ago
Unify word-level and span-level tasks: NJUNLP's Participation for the
WMT2023 Quality Estimation Shared Task
authored
a paper
6 days ago
Why Not Transform Chat Large Language Models to Non-English?
Organizations
None yet
models
7
DreamW1ngs/DCSQE-ende-word
Translation
•
Updated
•
2
DreamW1ngs/DCSQE-zhen-word
Translation
•
Updated
•
2
DreamW1ngs/DCSQE-zhen-sent
Translation
•
Updated
•
2
DreamW1ngs/DCSQE-ende-sent
Translation
•
Updated
•
2
DreamW1ngs/DCSQE-heen-syn
Translation
•
Updated
•
2
DreamW1ngs/DCSQE-zhen-syn
Translation
•
Updated
•
2
DreamW1ngs/DCSQE-ende-syn
Translation
•
Updated
•
2
datasets
0
None public yet