Notices by Kate Nyhan (kdnyhan@social.esmarconf.org)

@overholt
"A previous version of this article stated that the Harvard Slavery Remembrance Program had been disbanded. To clarify, according to a University spokesperson, the program has not been formally discontinued, though its employees and leaders have been laid off."

@overholt
How do people say that kind of thing with a straight face?

Guess what! The Brookline Public Library has CO2 monitors that you can borrow! Very useful for checking out the ventilation in any indoor space that is ventilated (mechanically or naturally) with outdoor air. FREE!!!!!
https://brookline.minlib.net/Record/.b42432650
#Brookline #BrooklineMA #IndoorAirQuality #COVIDisAirborne #CovodConscious #CovidCompetent

@anna have you heard about @fluconf ?

@ContraindiKate
Readimask - no metal nose piece (and no head straps either). It stays on with adhesive.

@ContraindiKate
https://readimask.com/
I was super excited when I saw the taupe mask the person is wearing in the pic on their website but it looks like they are selling only the yellow color....

@timnitGebru
About three hours elapsed between your post and my first opportunity to recommend the paper to a library user to whom it's very relevant :)

Can anyone recommend a guide about formatting a resume/CV for US federal govt jobs? My understanding is that it's pretty different from "normal"
#hiring

Machine indexing is going great, part infinity

"Biocultural and social determinants of ill health and early mortality in a New Mexican paediatric autopsy sample"

Subject headings include Mexico/epidemiology, but not United States or Southwestern United States or New Mexico.

Oh look

Title: A novel smart lighting clinical testbed
Abstract: A real-time, feedback-capable, variable spectrum lighting system was recently installed at the University of New Mexico Hospital to facilitate biomedical research on the health impacts of lighting. The system consists of variable spectrum troffers, color sensors, occupancy sensors, and computing and communication infrastructure, and is the only such clinical facility in the US.....

Indexing:

Come on, in this one the authors explicitly said right in the abstract, "the only such clinical facility in the US"!

That paper is from 2017 and I can no longer easily check machine vs human indexing status via Ovid Medline, so I don't know for sure if I should blame MTA or MTIA or MTIX or whatever it's currently called....

Title: Biocultural and social determinants of ill health and early mortality in a New Mexican paediatric autopsy sample
Indexing: Mexico / epidemiology

Title: Developing a Graduate Class on Synthetic Cells at a Minority Serving Institution: Lessons from the University of New Mexico
Indexing: Mexico

Abstract: "Data were collected from National Cancer Institute-designated cancer center at the University of New Mexico"
Indexing: Mexico

@mshoberecht
You want a scandal?
(Is it a scandal if the org blogs about it and nobody cares?)

MTIX, the new-and-improved version of machine indexing, the version that is *better* than the version used when NLM decided to switch from mostly-human to mostly-machine indexing in 2023, that MTIX...

is described *by NLM itself* as performing 66% as well as the original indexing method.

@mshoberecht
https://www.nlm.nih.gov/pubs/techbull/ma24/ma24_mtix.html

@mshoberecht
(And I would take them more seriously about speed as one of the goals of machine indexing if they ran PubMed records through MTIX to add subject headings *when the records were created in PubMed* instead of waiting for the record to be assigned an official publication date with a volume and issue.
As is, papers in Medline journals can languish for months or sometimes even years in "epub ahead of print" status without MTIX subject indexing :(

@mshoberecht
I am focusing on the "pther descriptors" category, not because I'm cherry-picking the lowest, but because other descriptors than check tags and publication types means -- the MeSH terms that #medlibs use for searching -- the headings like diseases and outcomes, populations that aren't checktags, research methods that aren't publication types....

@mshoberecht
And -- look, I'm just watching from the outside, maybe my guesses are wrong -- but we all know that named entity recognition for things like diseases can perform quite well. So if the disease terms are presumably being correctly applied to PubMed abstracts by MTIX, that suggests that the *non-disease* subject headings are being applied incorrectly, or missed when they should be applied, at a rate that is even worse than the 66% headline "other descriptors" score.

@mshoberecht
Hoorah, we are saving money and time and "overall" the quality is 74% as good as it used to be!

@mshoberecht
The National Library of Medicine's switch from largely-human to largely-machine subject indexing of biomedical journal articles is a great example of @pluralistic 's take, that it's a lot easier to persuade bosses to buy AI than it is to build AI that actually does the job.