An Assessment of the Impact of OCR Noise on Language Models
This addresses the problem of OCR noise degrading language model performance for researchers and practitioners working with digitized heritage texts, but it is incremental as it builds on existing assessments.
The study assessed how OCR noise affects various language models across Dutch, English, French, and German, finding that it significantly hinders performance, with simpler models like PPMI and Word2Vec outperforming transformers on small corpora.
Neural language models are the backbone of modern-day natural language processing applications. Their use on textual heritage collections which have undergone Optical Character Recognition (OCR) is therefore also increasing. Nevertheless, our understanding of the impact OCR noise could have on language models is still limited. We perform an assessment of the impact OCR noise has on a variety of language models, using data in Dutch, English, French and German. We find that OCR noise poses a significant obstacle to language modelling, with language models increasingly diverging from their noiseless targets as OCR quality lowers. In the presence of small corpora, simpler models including PPMI and Word2Vec consistently outperform transformer-based models in this respect.