[BTeam] LLMs to Extract Data from Logbook Entries

[Eduard Pozdeyev]

Thank you, Chris, for following up on this. The use of LLM to process available data can provide new opportunities, e.g., developing a smart ELOG search on details that are normally not searched or developing the Crew Chief/Operator Assistant. The system can also collect and process information on status of systems in the grey area (not directly monitored), like Facilities.

It makes sense to refine requirements and collect input from the operations, by collecting information on the needs. As with any new technology, people might think in term of already available capabilities. They need to be shown what new, out of the box capabilities might be available.

Eduard.

--
Eduard Pozdeyev
Director of Accelerator Operations
(office) 757-269-5770
[signature_299478068]


From: BTeam <bteam-bounces at jlab.org> on behalf of Christopher Tennant via BTeam <bteam at jlab.org>
Date: Wednesday, January 29, 2025 at 5:40 PM
To: bteam at jlab.org <bteam at jlab.org>
Subject: [BTeam] LLMs to Extract Data from Logbook Entries
Hi all,

I'm sending a link to a recent TN that uses large language models to extract data from images and tables in logbook entries. Some of you may be interested in the methodology while others may be interested in the dataset of 1,860 emittance measurements extracted from the last 9 years. Both can be found here:

https://jlabdoc.jlab.org/docushare/dsweb/View/Collection-59600

If there are other types of logbook entries that could benefit from this kind of analysis, please let me know.

Thanks,
chris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/bteam/attachments/20250212/515ef255/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 5149 bytes
Desc: image001.png
URL: <https://mailman.jlab.org/pipermail/bteam/attachments/20250212/515ef255/attachment-0001.png>