Why are AI learning models struggling to read older archival material?
Short answerBecause both the handwriting and the English differ from modern examples, so a model must learn period language as well as letterforms.
NotesA viewer reports that an archive's attempt to use AI on older documents is struggling. Older handwriting presents one difficulty, while changes in English create another. Recognizing the letters does not automatically supply the language knowledge needed to interpret them. Training a model on the English of the relevant period is proposed as part of the solution.