Meeting 15.4.2021

General

Next UD release is coming in the end of the month
Both treebanks have grown significantly,
Small fragment of Vászolyi stuff going to next UD release
- These are linguistically interesting fragments from prose version of the text. Some parts of the song could also be taken in, but that has some additional questions that may be better solved later.
Maria Fedina wants access to whole corpus in May-June: Should be done in connection to actually available release
Metadata needs to be checked: Especially the publicly visible part
Komi ASR model is part of Coqui ASR Model collection: https://coqui.ai/models, basically everything done by Inclusive Technology for Marginalised Languages group in Indiana is there. Should be documented better, and the role of our project clarified, but this is a very good development.
Niko has been in touch with Kielipankki and Aalto about their ASR tools, and we planned to add into their aligner Komi and Karelian options (basically the model does it with multilingual models, with best result probably for Finnish)

Niko is doing very large scale last-step bibliography check for both Udora and prolative papers, mainly to ensure there are no significant references that are currently being missed, and everything is certainly understood correctly
- So basically this means that every citation is verified again, and their sources are also traced back a bit
Niko and Jack were branstorming yesterday the IWCLUL paper about Komi dialectal FST stuff, this would connect well to the project's goals
Niko is planning to do something again with KyungTae Lim and Thierry Poibeau once PhD papers are on their way + Komi treebanks are in size level that makes their use relevant
Niko and Micha could go to the National Archives even during the corona restrictions, as long as the date is fixed well in advance, info here
- Suggestions about a possible date that works?
- This connects well to work done with Batalova materials