Dear list members, apologies for my ignorance, but where can I find an overview about the minimal requirements to set up a new language pair for Apertium MT? I'm considering to contribute to the development of translation system from Sumerian to English to complement our NMT (and EBMT) baselines ( https://github.com/cdli-gh), possibly in the context of the Apertium ecosystem.
Also, as we do have lexical resources and core components for morphology and syntactic parsing for Sumerian in place, I was also wondering about suggesting work on such a prototype as a topic for a GSoC project for 2021 -- and if there is any interest from the Apertium community in this, possibly as a collaborative effort (meaning joint mentorship) between the Cuneiform DIgital Library Initiative and the Apertium community. Clearly, a disadvantage of any MT system involving Sumerian is that the translation bridges languages that are not closely related by any means, but nevertheless, the relative data sparsity (this is an ancient language after all, even though well-documented) calls to explore directions of symbolic translation as yet another option, at least. Best regards, Christian
_______________________________________________ Apertium-stuff mailing list Apertium-stuff@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/apertium-stuff
