davanstrien
commented
3 years ago This is quite an interesting approach to deciding 'what is a worthwhile label':
Pre-trained machine learning models have become more widely available for commercial use over the past decade. Regarding models specifically for computer vision, many of the big players in the tech industry now offer services that automatically tag images with descriptive labels. We wanted to know whether it would be worthwhile to apply these tags to our collections in the new digital collections repository.
Once we stated this as our question, we were faced with another one: how would we measure improvement of metadata? We are a group of software developers, so we are familiar with user testing as a means of evaluating systems. We assert that if the users show increased satisfaction, then the system has been improved.
https://journal.code4lib.org/articles/15209