From Ledger to Query: OCR, NER, Search | Polay, Pemberton, Pasker | AI and History Conference 2026

Jacob Polay, Keith Pemberton and Bob Pasker on OCR, named entity recognition and machine-mediated search at the AI and History Conference 2026. Three speakers follow one pipeline from page image to query. Jacob Polay compares vision language model optical character recognition (OCR) with Tesseract on early modern and 19th century documents and proposes a tiered workflow that runs on a consumer GPU. Bob Pasker shows why keyword search misses antebellum testimony by Black witnesses and presents Roscoe, a conceptual search index over antebellum court opinions. Keith Pemberton argues that named entity recognition (NER) can only identify a name after it has been read, so it inherits the errors of OCR. Speakers: Jacob Polay, University of Saskatchewan. Keith Pemberton, Yale University. Bob Pasker, CUNY Graduate Center. 0:00 Welcome and Jacob Polay on OCR as the first step of the pipeline 1:56 A 1700 sugar plum broadside tested on six OCR systems 5:18 How vision language models differ from Tesseract and cut error rates 7:30 Plausible errors: Wooten for Westbury, Sioux Lake, rice turned into rye 11:41 Choosing an OCR model: why public leaderboards mislead 14:21 A tiered workflow on one NVIDIA 5070 Ti with Tesseract character counts 17:19 Bob Pasker on archival silences and why keyword search fails 24:16 Roscoe: conceptual search over antebellum court opinions 29:35 Build time, validation results and the cases Roscoe missed 41:13 Keith Pemberton: reading before naming, NER and its debt to OCR 47:32 A tiered routing pipeline that escalates uncertain names and keeps provenance 55:35 Questions on small OCR models, gold pages and kinship structures Talk page: https://proflouishyman.github.io/ai_c... Conference program: https://proflouishyman.github.io/ai_c... All talks:    • AI and History Conference 2026   Recorded at the AI and History Conference 2026, Johns Hopkins University (SNF Agora Institute), in partnership with the American Historical Association. Hosted by the Center for Economy and Society. #DigitalHistory #AIandHistory #HistoryAndAI #OCR #NER