The problem we are solving is fairly complex, and it would be useful to define more precisely what experimental work we intend to do, and how the models, datasets, and metrics link up into a consistent framework.
I've added a notes document and description of how to compile it.
This issue tracks keeping our notes up to date with latest discussion.
The problem we are solving is fairly complex, and it would be useful to define more precisely what experimental work we intend to do, and how the models, datasets, and metrics link up into a consistent framework.
I've added a notes document and description of how to compile it.
This issue tracks keeping our notes up to date with latest discussion.