training data, evaluation scripts and baselines released

[Andrey Kutuzov]

Dear fellow axolotl lovers,

This is an update from the organizers of the AXOLOTL-24 shared task on
explainable semantic change modeling
(https://github.com/ltgoslo/axolotl24_shared_task).

1) As planned, we are now publishing the training and development
datasets for Finnish and Russian (remember there will also be a third
surprise language at the test stage):
https://github.com/ltgoslo/axolotl24_shared_task/tree/main/data

2) We have also prepared a human-friendly explanation of the shared task
essence in one simple English example. If it was not entirely clear what
exactly we ask the participants to do, please have a look:
https://github.com/ltgoslo/axolotl24_shared_task/blob/main/example.md

3) The official scoring scripts for Track 1 and Track 2 are now
available, so you can immediately start evaluating your systems:
https://github.com/ltgoslo/axolotl24_shared_task/tree/main/code/evaluation

4) Finally, baseline systems are prepared for both tracks, so that you
can have something to outperform:
https://github.com/ltgoslo/axolotl24_shared_task/tree/main/code/baselines

The next step in our timeline is March 25, 2024 when the test data will
be published.

If you have any questions, do not hesitate to ask - in this mailing list
or in Github issues.


--
Andrey
Language Technology Group (LTG)
University of Oslo