Unverified50% confidenceBenchmarkExact time
On all low-resource languages, back-translation, compared with fine-tuning on parallel data alone, improves the XCOMET-XXL score by 16% to 173% in the direction of translation into the low-resource language.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOn Odia and Swahili, the back-translation method, using only 1000 parallel sentence pairs, outperforms continued pre-training by 7.7 to 16.4 XCOMET-XXL points in in-domain evaluation.70% similarUnverifiedBenchmark datasets for sign language translation (such as PHOENIX-2014T, CSL-Daily, and How2Sign) typically contain only thousands to tens of thousands of samples, far smaller than the millions of sentence pairs in parallel corpora in the machine translation field.66% similarUnverifiedFor low-resource language translation, the most effective way to use monolingual data is not continued pre-training, but back-translation.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1028056API
curl https://kongchang.com/api/v1/knowledge/claims/1028056MCP
get_claim(id=1028056)