Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

scrollprize.org

Post reply on HN
Print view View original

101–110 of 218 posts

Re: Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

#101

Post by saled »

linsomniac wrote: post #88
alach11 wrote: post #4

This is the coolest thing I've read this year. It reads like science fiction. Who would even imagine it's possible to read text from a 2000 year-old rolled up burnt-crisp paper?

Speaking of "reading like sci-fi", what's that book where they scan an entire library of books, descructively, by feeding them into a "book chipper" like device that chops the books up into little pieces, vacuums those pieces up and scans the pieces as they flow through, reconstructing the original text by putting the scanned results together like so many jigsaw puzzles? It was a subplot of the book, but I can't for…

FYI it seems ChatGPT could have answered this for you.

> The book you're describing sounds like "Rainbows End" by Vernor Vinge. In this near-future sci-fi novel, set in 2025, one of the subplots involves a project called the "Library Project," where the UCSD (University of California, San Diego) library decides to digitize its entire collection. The process is somewhat as you described: books are destructively scanned by being shredded into tiny pieces, which are then scanned and digitized, with the text being reconstructed from the scans. This process is a part of the broader themes of the book, which include the effects of technology on society and the concept of "wearable computing" and augmented reality. Vernor Vinge, a retired San Diego State University professor of mathematics, computer scientist, and Hugo Award-winning author, is well-known for his works in the science fiction genre, especially for exploring the concept of the technological singularity.

Re: Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

#102

Post by gwern »

saled wrote: post #101
linsomniac wrote: post #88

Earlier quoted context omitted.

Speaking of "reading like sci-fi", what's that book where they scan an entire library of books, descructively, by feeding them into a "book chipper" like device that chops the books up into little pieces, vacuums those pieces up and scans the pieces as they flow through, reconstructing the original text by putting the scanned results together like so many jigsaw puzzles? It was a subplot of the book, but I can't for…

FYI it seems ChatGPT could have answered this for you. > The book you're describing sounds like "Rainbows End" by Vernor Vinge. In this near-future sci-fi novel, set in 2025, one of the subplots involves a project called the "Library Project," where the UCSD (University of California, San Diego) library decides to digitize its entire collection. The process is somewhat as you described: books are destructively scanne…

I'm not surprised ChatGPT can answer it - I'm not sure why, but _Rainbows End_ is one of the most commonly-asked about SF books like that. Everyone remembers the book-tornado doing shotgun sequencing, but they can never remember its name or anything else that happens. I guess that's the problem with having a technology whose mental image is so compelling but also mostly disconnected to the rest of the book. (I know I can't tell you much about the rest without rereading the WP entry.)

Re: Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

#103

Post by gwern »

codeulike wrote: post #18

Over 2000 years ago a chap called Philodemus sits in the library of a luxurious villa owned by a rich guy who likes collecting art and writing. He writes his thoughts on pleasure, and the relationship between the quantity of something and the pleasure that might derive from it. The scroll goes on the shelves with the others. He writes lots. At some point later the villa is covered by lava from mt vesuvius. 2000 years…

Epicureans were atomists (https://en.wikipedia.org/wiki/Epicureanism#Physics), so if you explained it to him, he might not be nearly as shocked as most people of the era would be. Epicureans also tended to invoke extremely insubstantial 'images' composed of especially tiny gossamer arrangements of atoms as explanations for things like dreams, and you can see how well that would work as an analogy with using X-rays to look at subtle changes in the arrangement of atoms in charred scrolls. (These are covered in https://en.wikipedia.org/wiki/De_rerum_natura - if you have some time, I highly recommend reading Stalling's rhyming translation. You'll be shocked and admire the rationality & scientificness of Epicurean materialist atomist explanations of the world, even where they get it totally wrong.)

Re: Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

#104

Post by kortex »

jtchang wrote: post #6

“Any sufficiently advanced technology is indistinguishable from magic.” Absolutely insane the level of wizardry being applied here to turn a lump of blackened, charred scrolls into readable text. Having only cursory knowledge with machine learning are some of the techniques used in the article only recently discovered or have they been around for a while? Is it due to us having reached an inflection point with these…

There has definitely been a virtuous cycle between GP-GPU processing capability, algorithms, libraries and software that use that hardware, and researchers working with those tools.

Re: Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

#105

Post by cameron_b »

_a_a_a_ wrote: post #17
dmurray wrote: post #11

Earlier quoted context omitted.

An example of the same thing at a macro level: https://www.smithsonianmag.com/smart-news/archaeologists-reb...

Bigger yet by far https://en.wikipedia.org/wiki/Mausoleum_of_the_First_Qin_Emp... He of the terracotta army. Not excavated yet for fear of damage, but I would so love to know...

The feeling you get when you’ve gone into one of those aircraft hanger-size buildings and then you see some of the information they’ve gotten with ground tests ( radar, mercury, etc ) is wild. The site is huge.

One of the suppositions is that the main chamber contained a model of his entire kingdom, replete with rivers of mercury.

So yes. Archaeology is a bit destructive, and sometimes the destruction can go both ways. Proceed with caution.

Re: Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

#106

Post by p-e-w »

kretaceous wrote: post #5

When I first came across this project on HN (early last year), I was taken aback by how impossible the project looked and how smart were people working on this. Despite seeing a few intelligent names behind the project, I subconsciously believed that this would at least take 5-10 years before a breakthrough. Today I sit with the same amazement, taken aback again, appreciating how ridiculously awesome this is. Congrat…

It's all about incentives. $1 million is a lot of money. The vast majority of hard problems don't have much brainpower dedicated to them, because the bang/buck ratio doesn't work out. Machine learning, math, and adjacent fields already have many careers that pay very well, so getting top-notch experts to dedicate their attention to what might be a futile endeavor is difficult.

And this isn't only about the monetary value itself, but also the fact that a large cash prize attached to a challenge boosts the prestige of finding a solution. Nobel Prizes come with about a million bucks on top of them, after all.

I'm quite confident that if someone offered $100 million for deciphering the Voynich manuscript or Linear A, we'd have a solution within 3 years.

Re: Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

#107

Post by BenFranklin100 »

p-e-w wrote: post #106
kretaceous wrote: post #5

When I first came across this project on HN (early last year), I was taken aback by how impossible the project looked and how smart were people working on this. Despite seeing a few intelligent names behind the project, I subconsciously believed that this would at least take 5-10 years before a breakthrough. Today I sit with the same amazement, taken aback again, appreciating how ridiculously awesome this is. Congrat…

It's all about incentives. $1 million is a lot of money. The vast majority of hard problems don't have much brainpower dedicated to them, because the bang/buck ratio doesn't work out. Machine learning, math, and adjacent fields already have many careers that pay very well, so getting top-notch experts to dedicate their attention to what might be a futile endeavor is difficult. And this isn't only about the monetary v…

Not to be argumentative, but $1M isn’t very much money, certainly not for a project of this scope. It’s a testament to the creativity, competence, and dedication of those involved they’ve gotten this far with such little funding. Hopefully their early success will attract more resources to this very worthy project.

Re: Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

#108

Post by mNovak »

The scroll segmentation looks like it'd be a very manageable task for distributed volunteer work too, at least as a kickstart if that's really the bottleneck now. Just a 100 or 1000 halfway dedicated volunteers (small by internet standards, for a useful and simple task) could make a big dent. Don't know how many scan layers are in a single scroll, but for instance Project Gutenberg's proofreading network has processed millions of pages, one by one.

Re: Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

#109

Post by CityOfThrowaway »

p-e-w wrote: post #106
kretaceous wrote: post #5

When I first came across this project on HN (early last year), I was taken aback by how impossible the project looked and how smart were people working on this. Despite seeing a few intelligent names behind the project, I subconsciously believed that this would at least take 5-10 years before a breakthrough. Today I sit with the same amazement, taken aback again, appreciating how ridiculously awesome this is. Congrat…

It's all about incentives. $1 million is a lot of money. The vast majority of hard problems don't have much brainpower dedicated to them, because the bang/buck ratio doesn't work out. Machine learning, math, and adjacent fields already have many careers that pay very well, so getting top-notch experts to dedicate their attention to what might be a futile endeavor is difficult. And this isn't only about the monetary v…

I'm 90% sure the people that did this project did it because they got nerd sniped by it and got to hang out with nat while earning a reasonable salary

Re: Vesuvius Challenge 2023 Grand Prize awarded: we can read the first scroll

#110

Post by fragmede »

BenFranklin100 wrote: post #107
p-e-w wrote: post #106

Earlier quoted context omitted.

It's all about incentives. $1 million is a lot of money. The vast majority of hard problems don't have much brainpower dedicated to them, because the bang/buck ratio doesn't work out. Machine learning, math, and adjacent fields already have many careers that pay very well, so getting top-notch experts to dedicate their attention to what might be a futile endeavor is difficult. And this isn't only about the monetary v…

Not to be argumentative, but $1M isn’t very much money, certainly not for a project of this scope. It’s a testament to the creativity, competence, and dedication of those involved they’ve gotten this far with such little funding. Hopefully their early success will attract more resources to this very worthy project.

it's $700k divided three ways, too. $234k is well within FAANG compensation range, but you get to work on such an awesome project.
Post reply on HN