Unlocking New Secrets: Rediscovering Historical Documents with AI
The Louis-François-Georges Baby Collection of historical documents, preserved by Université de Montréal’s Archives and Information Management Division (DAGI), is a treasure trove of information about New France’s history. Now, thanks to the efforts of Jean-Olivier Dicaire-Leduc, a master’s degree student in history at UdeM, the collection is revealing its secrets with the help of artificial intelligence (AI). Dicaire-Leduc utilized Transkribus, an open-source software developed by the University of Innsbruck, Austria, to decipher the handwritten documents found in the collection.
The N Series of the Baby Collection contains a variety of handwritten documents related to aboriginal affairs, but many of them are illegible. Utilizing Transkribus, Dicaire-Leduc scanned high-resolution digital images of the manuscripts and fed them into the software, which automatically transcribed the handwritten text. The software uses learning algorithms to identify recurring words and turns of phrase, deciphering the message within the documents.
However, the automated transcriptions generated by Transkribus were not perfect, containing some errors. Dicaire-Leduc, therefore, had to manually correct the transcriptions to enhance readability and improve indexing. While the process was challenging due to the diverse subjects and authors covered in the N Series documents, it allowed Dicaire-Leduc to gain insights into a crucial part of New France’s history.
Dicaire-Leduc had three main goals for his internship with DAGI. First, he aimed to transcribe the documents to enhance accessibility and facilitate indexing using UdeM’s AtoM archival description app. Second, he sought to contribute to UdeM’s initiative to promote archiving in the digital age by sharing historical and archaeological data related to the people living in the Montreal area during the 17th century. Lastly, Dicaire-Leduc aimed to critically review the archival descriptions produced when the Baby Collection was initially archived 70 years ago, providing context without eliminating terms that have aged poorly.
The Louis-François-Georges Baby Collection was amassed by Louis-François-Georges Baby, a prominent figure in Canadian history. Baby, who served as the mayor of Joliette and a federal MP for Joliette, collected over 20,000 archival documents spanning three centuries. The collection covers diverse subjects such as agriculture, education, military affairs, literature, and politics. Notable items include documents signed by historical figures like King Louis XIV of France and the Cardinal de Richelieu.
Dicaire-Leduc’s successful use of AI to unlock the secrets of the Baby Collection demonstrates the potential of technology in historical research and archiving. By combining AI with human correction and analysis, researchers can make historical documents more accessible and gain new insights into the past. The work done by Dicaire-Leduc has the potential to create research aids and dissemination tools that benefit historians and the public alike, facilitating a deeper understanding of New France’s history.