Combining a diffusion model with a sparse coding strategy to improve the machine translation accuracy of culturally loaded words in English news texts
Qin Yu, Tao Mai · 2026
Read Combining a diffusion model with a sparse coding strategy to improve the machine translation accuracy of culturally loaded words in English news texts by Qin Yu, Tao Mai (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0356146.txt.
Description
Qin Yu and Tao Mai build a translation system aimed at the hardest words in English news. The targets are culturally loaded terms, metaphors, and institutional names. Such items lack support in training data. Such long-tail items drift in meaning or lose their cultural charge in standard machine translation. Their design first squeezes text into a sparse code. The code spotlights cultural elements. Then it rebuilds the translation step by step. A diffusion process does the rebuilding, guided by gates and a cultural knowledge base.
On long news sentences, the system beat three mainstream rivals on two standard quality scores. On unseen vocabulary, it kept accuracy and recall well above the baselines. So the authors report. Yu and Mai present the hybrid as a route to translations that carry culture, not just words.
Catalog record
| CREATOR | Qin Yu, Tao Mai |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 76286 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |