To improve template support a machine learning approach (T221211) was applied to identify the mappings/alignments of parameters for the most used templates and language pairs (generated alignments). We want to generate additional mappings of Urdu, Kurdish, Igbo, Haitian Creole, Czech, Slovak, Turkish, Xhosa, Zulu, Shona, Luganda, Hausa, Occitan, Yoruba, Bashkir, Azerbaijani, and Malay.
In particular, we want support these additional pairs:
Run 1:
- en → az
- en → ba
- en → cs (Current mapping: 243)
- en → ha
- en → ht
- en → ig
- en → ku
- en → lg
- en → ms
- en → oc
- en → sk
- en → sn
- en → tr (Current mapping: 387)
- en → ur (Current mapping: 82)
- en → xh
- en → yo
- en → zu
Run 2:
- ca → oc
- cs → sk
- fr → ht
- fr → oc
- ru → ba
- sk → cs
Run 3:
- simple → xh
These are languages used on wikis that may be interested in Section Translation. The set includes the target wikis where the Growth team is deploying their recent tools (which may integrate with Section Translation in the future).