New tools for old news

Abstract

Many collections of digitized newspapers suffer from poor OCR quality, which impacts readability, information retrieval, and analysis of the material. Errors in OCR output can be reduced by applying machine translation models to “translate” it into a corrected version. Although transformer models show promising results in post-OCR correction and related tasks in other languages, they have not yet been used for correcting OCR errors in Swedish texts. This thesis presents a post-OCR correction model for Swedish 19th and 20th century newspapers based on the pre-trained transformer model ByT5. Three versions of the model were trained on different mixes of training data. The best model, which achieved a 37% reduction in CER, will be integrated in Språkbanken Text’s annotation pipeline Sparv.