“Using AI to Transcribe, Summarize, and Extract Information from Handwritten Genealogical Documents”
Dallan Quass
7/13/24
Categories: IR, RT
Talk Abstract:
I will explain how to use multi-modal large language models (LLMs) to transcribe handwritten genealogical documents and also how to use them to summarize the documents including names, dates, and places in the summaries. I will compare the cost and performance of several LLMs so you can decide which one is best for you.
Speaker Bio:
I have a PhD in computer science from Stanford University and BS and MS degrees in computer science from BYU along with extensive software development experience, especially extracting information from the Web and creating leading-edge search engines. I have served as the CTO of FamilySearch and have co-founded several software startups: RootsFinder, acquired by Findmypast, WhizBang! Labs, acquired by Inxight Software, FlipDog, acquired by Monster.com, and Junglee, acquired by Amazon.com. My interest in family history prompted me to develop two other family history websites: gengophers.com and werelate.org, both of which have been listed among the top 101 family history websites according to Family Tree Magazine. Finally, I’ve written over a dozen academic publications which have been cited over 6,000 times according to academic.research.microsoft.com.
Viewing of videos and handouts is only available to members who have logged in. If you are a member, please login or sign up using the login button in the upper right corner of this page.
NOTICE:
This presentation is part of a set of over 400 presentations on genealogy and family history produced by "UVTAGG: The Utah Valley Technology and Genealogy Group".
For full details and to join, see the website https://uvtagg.org.