I study how machines read imperfect traces.
My work is on AI for noisy documents: historical newspapers, OCR and HTR transcripts, handwritten archives, inscriptions, administrative and financial records. I build methods to extract people, places, events and relations from them, and I measure what models miss, invent or flatten along the way.
La Roche1le, l4 juillet 189O. — Le Conscil municipal s’est réuni à l’Hôtel de Vi11e.
La Rochelle, 14 juillet 1890. — Le Conseil municipal s’est réuni à l’Hôtel de Ville.
Six threads I keep returning to
Each thread links to a page with the problem, the approach, and the papers behind it.
Noisy documents
OCR and HTR errors, degraded text, and how noise changes downstream meaning, not just character accuracy.
Research + papers →Semantic extraction
Named entities, entity linking, relations and events: the small structures that make collections searchable.
Research + papers →Historical AI
Temporal modelling and multilingual archives, for language whose names, borders and meanings keep shifting.
Research + papers →Cultural heritage
Digital epigraphy, Armenian and Ukrainian inscriptions, structured vocabularies, and contested memory.
Research + papers →Multimodality
Text, images, layout and visual signals modelled together, rather than treating everything as plain text.
Research + papers →Applied document intelligence
Document fraud, misinformation, epidemic surveillance, and open LLMs built to hold up under real constraints.
Research + papers →Recent and selected papers
1,180 citations · h-index 17 on Google Scholar. Below, all my publications by research theme and year.
Notes and essays
Shorter texts on AI, archives, evidence and noise. Less formal than papers.
- Essay
LLMs in archives: fluent, useful, suspicious
Why good prose is not the same thing as grounded historical understanding.
- Note
Documents are evidence, not just data
Building AI systems that keep traces attached to claims.
- Note
Noise is not always a bug
OCR, damaged scans, and why cleaning can also erase what matters.
Mostly serious work, with a sense of humour
I care about what AI extracts from documents, and also about what it misses, invents, flattens, or quietly ruins while sounding confident.
I come from both computer science and visual practice, so I tend to look at documents as text, images, layouts and evidence all at once. My PhD was on neural methods for event extraction; since then I have worked on historical newspapers, epigraphy, document fraud, epidemic surveillance and open LLMs.
Outside research: photography, handmade things, gardening with variable success, mostly non-fantasy books, metal festivals and camping.
- FocusNLP, information extraction, document AI, digital humanities
- PhDNeural Methods for Event Extraction (2018)
- LinksGitHub · Google Scholar · ORCID · LinkedIn