Needs more information
- What are you trying to accomplish? Why?
- Use Large Language Models and ai to search for sources that can provide more information. Document the source and attach to the individuals in question
- What is the challenge or roadblock you are encountering?
- Completion of the temple ordinance because of lack of information
- What is your idea?
- Use LLMs to index the web for source information about an deceased individual
- How would solving this challenge improve your experience?"
- It would reduce time to complete the temple ordinance
Comments
-
I'm sorry, but any mention of automation of the processes on FamilySearch gives me night terrors. You will find many threads giving examples of the issues created by poorly-written algorithms.
1 -
The idea of a list of additional source candidates (presumably for manual verification before inclusion as sources in FT) is definitely interesting. I have been analysing downloaded (TSV) Records metadata via SQL with a similar sort of end in view, with some success, although by definition I can only analyse the metadata that it has occurred to me to download.
Maybe your approach could be applied to the FS Catalog, given that it is already able to search for useful information within online books (even those to which the logged in user doesn't have full access). Also, I find it very frustrating that FS has many unindexed Images containing typewritten or computer-generated text which my PC can read (once I download them) but for which FS has no 'full text search' capability at all - maybe this idea could help with these, too.
But if you are talking about sources that are hard to find through 'normal' FS and other genealogical research routes, for example those that are currently locked in unindexed (or incorrectly indexed) handwritten images, I'm not sure how much benefit we'd get in 2024 (though obviously all bets are off re future developments).
0
