The title calling this a "material" is disingenuous. One would assume that they somehow tested a "bulk" material coupon even if just a few mm^2. But the image in the article shows them testing a single knot.
Funny thing is the article mentions machine learning as one of the consumers of the digitized books. Maybe some of the money going to machine learning now may actually contribute to preserving original source material.
In archaeological terms, they could be discovering stuff like microscopic grain husks embedded into the clay of the containers. Though they could have found whole chicken skeletons or fossilized garlic or something else too... I searched for a moment but couldn't find more details from this dig about the food specifically.
They describe the modern technique in the UPenn announcement...
Rather than digging according to architectural construction phases, the Lagash Archaeological Project is using an approach championed by Pisa’s Pizzimenti, who excavates by microstratigraphic layers, thin lens by thin lens horizontally, across a wide swath, “like doing very careful surgery,” Pittman says. “Just 50 centimeters down, we were able to capture all of this. We were happily astounded.”
Get tested for diabetes just in case? Urination and dry mouth are symptoms that I see you mentioned which don't seem to be apnea related. I'm not a doctor.
Usage notes
Meteor (streak of light in night sky): Not to be confused with meteoroid and meteorite (cause and remains of a meteor), or asteroid and comet (celestial bodies).
To do that with the current level of "learning" I think you will need a training set that has lots of "comprehension"... Maybe a bunch of meta-analysis papers, UN summary reports. Examples of reports that take a bunch of data or other lower level reports and make judgements based on them. Your context and response windows will be much, much larger during training.