Welcome To the Bonny Broom Blog
Welcome to the Bonny Broom Research blog! Here we’ll share updates about Videlicet™, research highlights, and news about our work making history searchable.
My last semester as an undergraduate is officially over! Finals have been turned in, grades posted, and grad caps thrown. Videlicet is up and running, and I’m excited to continue developing this project during the summer.
I wanted to bring you up to speed on what Videlicet is and where it’s going.
Videlicet is a digital humanities research project created to address the needs of contemporary historical research. Away from well-known documents like the Declaration of Independence or Federalist Papers, many historical artifacts are not easily readable, transcribable, or searchable. The technology used by all major historical databases, Optical Character Recognition (OCR) and keyword searching, is often not compatible with the idiosyncrasies of older documents. I use exciting developments in AI technology to address these shortcomings.
Videlicet is an ongoing, iterative project that trains Large Language Models (LLMs) to transcribe historical documents.
There are two components of the project that make Videlicet innovative and exciting:
-
Patent-pending accuracy measurement system. During data intake, select documents are transcribed by people and compared to computer-generated transcriptions. Researchers evaluate the results to improve computer accuracy. AI transcribing should never be a one-and-done action! The process can always be improved with your input and critical eye.
-
Patented semantic search. Videlicet does not just generate document transcriptions. It provides a whole search engine too! Our patented semantic search system identifies word meaning, rather than character recognition, to trace trends and identify relationships between documents. People in the 1600s were not trying to make their writing word-searchable for online databases! Semantic search accommodates for multiple spellings or meanings of a single word.
I’ve included a handy visual to show you how our system works. Our process is intuitive, transparent, and flexible with your research or archiving needs.
I’ll be posting updates for our exciting system! Follow along to catch what we’re up to and discover how we can help you make your historical documents more searchable.
