https://www.livescience.com/ai-is-deciphering-a-2000-year-old-lost-book-describing-life-after-alexander-the-great
Machine learning and computed tomography scans being used to read ink traces on a carbonized scroll buried in the Vesuvius eruption.
QuoteOnly small parts of the heavily damaged text can be read right now. "It contains the names of a number of Macedonian dynasts and generals of Alexander," Janko said, noting that it also includes "several mentions of Alexander himself." After Alexander the Great died in 323 B.C., his empire fell apart. The text mentions the Macedonian generals Seleucus, who came to rule a large amount of territory in the Middle East, and Cassander, who ruled Greece after Alexander's death.
Amazing stuff. I wonder what we will learn.
I am getting really excited about this: I wonder what (if any) new information, or confirmation will will get.
I wonder when we will get the first texts?
Wow! This is absolutely fascinating. I doubt the idea that it was from a book 'borrowed' from elsewhere, however. The owner of the Villa of the Papyri had a large private library (the contents of which are gradually being uncovered bit by bit). There's a good documentary about it here: https://www.youtube.com/watch?v=98W-j545-0Y
The library included a lot of philosophical works, especially by Philodemus. But just imagine this. The villa has revealed several busts identified as Alexander's successors (Seleucus, Lysimachus, Demetrius). There are also a few frescoes based only early Hellenistic originals, perhaps even a portrait of a Macedonian king or two. Therefore, might it be possible that the owner extended his interest to include books on the same personalities / historical period? I've been keeping my eye on the discoveries in that villa for quite some time, and often dreamed of there being a lost copy of Hieronymus of Cardia buried and carbonised there...and awaiting discovery!
With a career spent as a math/stats fellow in data science, I enjoy seeing machine learning techniques applied to obscure and long puzzling challenges. Having used machine learning to reduce the information set in an attempt to identify different types of imagery for classification and decision-making, I will be following this project with great interest.
Thanks for bringing this project to our attention.
pretty clever stuff
There is more (but slightly older) detail on the techniques on the project website (https://www2.cs.uky.edu/dri/herculaneum-papyrus-scrolls/), with a collection of videos (https://www2.cs.uky.edu/dri/videos/).
Personally I'm less hopeful of a complete text being recovered rather than just words or fragments but who knows - just 20 years ago even this much would have been impossible.
Quote from: RichT on February 16, 2023, 10:35:16 AM
There is more (but slightly older) detail on the techniques on the project website (https://www2.cs.uky.edu/dri/herculaneum-papyrus-scrolls/), with a collection of videos (https://www2.cs.uky.edu/dri/videos/).
Personally I'm less hopeful of a complete text being recovered rather than just words or fragments but who knows - just 20 years ago even this much would have been impossible.
Agreed, but even fragments can offer unique information - it was the recent fragment of Dexippus that proved that Ostrogotha really did exist, when modern historians had been united in saying he was just a made up eponym. Here's hoping anyway!