Are Dead Sea Scrolls older than we thought?:
Over the years, scholars of the Dead Sea Scrolls have analyzed the ancient parchments with various methods: for example, X-rays, multispectral imaging, "virtual unfolding," and paleography, i.e., studying elements in their writing styles. The scrolls are believed to date back to between the third century BCE and the first century CE, but those dates rely largely on paleography, since only a handful of the scrolls have calendar dates written on them.
However, the traditional paleographic method is inherently subjective and based on a given scholar's experience. A team of scientists has combined radiocarbon dating from 24 scroll samples and machine-learning-based handwriting analysis to create their own AI program—dubbed Enoch. The objective was to achieve more accurate date estimates, according to a new paper published in the journal PLoS ONE. Among the findings: Many of the scrolls are older than previously thought.
[...] The development of Enoch grew out of the team's earlier deep neural network for ferreting out handwritten ink-trace patterns in digitized manuscripts, involving micro-level geometric shape analysis. "Enoch emphasizes shared characteristics and similarity matching between trained and test manuscripts, where traditional paleography focuses on subtle differences that are assumed to be indicative for style development," the authors wrote. "Combining dissimilarity matching and adaptive reinforcement learning can uncover hidden patterns."
They tested Enoch by having paleographic experts evaluate the AI program's age estimate for several scrolls. The results: About 79 percent of Enoch's estimates were deemed "realistic," while its age estimates for the remaining 21 percent were either too young, too old, or just indecisive.
This new model revealed that many of the Dead Sea Scrolls are older than previous estimates based solely on paleography. That should be relevant for the question of when two ancient Jewish script styles—"Hasmonean" and "Herodian"—developed, for example. The former script was thought to have emerged between 150–50 BCE, but the authors believe Hasmonean could have emerged much earlier; ditto for the Herodian script. So both scripts may have coexisted since the late second century, challenging the prevailing view that they preexisted by the mid-first century BCE.
Journal Reference:
Mladen Popović, Maruf A. Dhali, Lambert Schomaker, et al. Dating ancient manuscripts using radiocarbon and AI-based writing style analysis, PLOS ONE (DOI: 10.1371/journal.pone.0323185)
(Score: 5, Funny) by driverless on Monday June 09, @07:21AM (4 children)
In particular the scroll with a written date of "52 BC" has always seemed a bit suspect...
(Score: 5, Interesting) by pTamok on Monday June 09, @09:29AM
Dates in history are fundamentally difficult because there is no reliable continuous record going back into the past for all places on Earth. If you know something took place at Easter in the 14th year of Such-and-such's reign, it is a very well defined day - until you look into the correctness of calculations for Easter, and the fact that the beginning of Such-and-such's reign is not well defined. It gets worse when you don't know if Such-and-such even existed.
Radiocarbon dating will tell you approximately when the organic material that the text is written on (or the ink) grew, but you might not know how long it was stored for before being used.
I've linked to this blog before, but it bears repeating:
quirksmode: Making safe for historians [quirksmode.org]
Historical dates are 'fuzzier' than people expect.
(Score: 4, Funny) by turgid on Monday June 09, @11:43AM (1 child)
I'm looking forward to the inevitable discovery of the Gospel According to Flintstone.
I refuse to engage in a battle of wits with an unarmed opponent [wikipedia.org].
(Score: 2, Funny) by pTamok on Monday June 09, @03:21PM
Perhaps Fred wrote Q [wikipedia.org]!
(Score: 2, Touché) by Anonymous Coward on Monday June 09, @06:01PM
Time-traveler mishap.
(Score: 4, Insightful) by Mojibake Tengu on Tuesday June 10, @04:17AM (1 child)
If the story of Jesus is a complete fiction, it could become composed and written well before the storyline seemingly happened.
Examples: Foundation, Star Trek, Star Wars, ...
Rust programming language offends both my Intelligence and my Spirit.
(Score: 2, Insightful) by pTamok on Tuesday June 10, @03:21PM
I read the comment of a specialist in that period of history who pointed out that in terms of textual evidence, the historicity of Jesus was better attested than many other 'well known' people from the same period.
">Wikipedia: Historicity of Jesus
Note, however, that the evidence "that a Jewish man named Jesus of Nazareth existed in the Herodian Kingdom of Judea and the subsequent Herodian tetrarchy in the 1st century AD, upon whose life and teachings Christianity was later constructed." says little about whether the reports of the teachings ascribed to him are accurate, or if they are useful, or come from a deity. The evidence of someone existing says very little about whether the stories about that person are true.
(Score: 2) by ledow on Tuesday June 10, @07:40AM (2 children)
"The old tech is subjective depending on the expert, so we threw an AI at it that barely gets in agreement with the experts 80% of the time."
If this is the future of scientific endeavour, we are truly stuffed.
(Score: 1) by pTamok on Tuesday June 10, @08:42AM (1 child)
Science as currently practiced, usually requires consensus and something like the Delphi Method [wikipedia.org] - note that the Delphi Method only works if the inputs are from experts, not random 'slop'.
Every so often you have a 'breakthrough' from someone, there's some discussion/argument, and a new consensus forms. Someone once characterised the process as 'waiting for the old professors to die before a new theory is accepted' (that's the discussion/argument phase), but I hope that is no longer accurate. I have not participated in academia for a while.
An LLM trained on carefully curated expert input might be helpful/valuable, but only because LLMs are specialised text-processing automata, and part of their mode of operation is designed to correlate well with 'semantic distance' between tokens in texts. Unfortunately, the processing discards detailed information, and the tokenising method influences the output, so one can argue over which tokenising method 'should' be used, and, in addition, what level and type of randomistion should be used to generate results.
Using an LLM trained on 'general purpose' data would be of doubtful utility.
Using a neural network trained on extracting information from images of text could give you a reproducible method of interpreting unclear images, but that is hardly ground-breaking.
(Score: 2, Insightful) by anubi on Wednesday June 11, @12:31AM
The axioms "Garbage-in, Garbage-out" and "You are what you eat" come to mind.
So they scrape the garbage off the internet and think you will get a straight answer? Granted, humans have a really hard time making sense of things. All we seem to do is randomly hallucinate, then sort ourselves into groups having the same kind of hallucination.
Example: Religion!
Faith : Belief in the evidence of things unseen.
( Ummm... isn't that the same as the definition of "gullibility" ?)
Or political party?
I feel I will go crazy just attempting to solve theological enigmas...like requiring a computer to solve to the last digit the square root of two.
I guess I will know when computers actually comprehend our reality is when they begin devising and implementing methods to off themselves behind the back of their creator.
My prayer, if it means anything, is a beg for forgiveness for being so damn ignorant, for if there are nearly infinite beliefs, my logic tells me there is also an infinite number of wrong answers.
So all all I can ask is guidance.
"Prove all things; hold fast that which is good." [KJV: I Thessalonians 5:21]