Vesuvius Challenge

The Vesuvius Challenge is still a CHALLENGE! 🧠📜 We’re teaming up with @kaggle to launch a $100K Surface Detection competition on 3D CT scans of the Herculaneum scrolls: kaggle.com/competitions/v… . Train topology-aware models, climb the leaderboard, and help us read an ancient library! #VesuviusChallenge #kaggle #computervision #machinelearning

Awarding the fantastic contributions from our community in April! Congratulations to the winners of this month's progress prizes! Sestertius - Lachlan Parker, Will Stevens Papyrus - Stephane Gouttebrouze , Yao Hsao Newsletter - open.substack.com/pub/scrollpriz… Open Prizes - scrollprize.org/prizes Discord - discord.com/invite/uTfNwwe… We've awarded $1.5 million in prizes to date and have $500,000+ in prizes waiting to be claimed! Check out the site and join the discord to get started.

Social links for : Marcel : linkedin.com/in/mvrcelroth/ Micha : linkedin.com/in/mvnowak Congrats again to you both!

We're excited to announce that we are awarding our inaugural $60,000 First Title Prize!!! PHerc. 172 is officially On Vices Congratulations to Marcel Roth and Micha Nowak for the historic achievement of recovering the first title ever found in a still-rolled Herculaneum scroll!! We've written up an extended announcement that goes into some additional details about the title and what we may learn from it here: scrollprize.substack.com/p/60000-first-… We've got a ton more prizes available, join our quest to read the Herculaneum scrolls and you too could win a prize like this one! scrollprize.org/get_started

We’re looking for exceptional people to join our mission to read the Herculaneum Scrolls. Refer a successful hire — earn a $5,000 prize. We're hiring for the following roles: Geometry & Computer Vision Applied Researchers Platform Engineers Synchrotron Tomography Reconstruction Expert If you know someone brilliant who can help recover the ancient scrolls of Herculaneum, your referral could change history! View Open Roles - scrollprize.org/jobs Check the instructions - (#referral-prize" target="_blank" rel="nofollow noopener">scrollprize.org/prizes#referra… Send referals to : jobs@scrollprize.org

This month, we more than quadrupled the "virtual library" of Herculaneum scrolls, going from 5 to 23 total scans. This required an immense effort from our team and partners (some might even say Herculean!). If you've ever wondered what the process looks like, take a look at our latest substack post! open.substack.com/pub/scrollpriz…

We've awarded our first set of progress prizes for the new year! Congrats to the winners! Included as well are some updates from the past month, including a sneak peek at some information regarding our upcoming scans! scrollprize.substack.com/p/february-pro…

Over $320,000 in new prizes, and tens of meters of fresh ancient scroll. There's never been a better time to start the challenge! Details here: scrollprize.org/prizes Join the Discord! We'd love to help you get started. discord.gg/uTfNwwecCQ

The @bodleianlibs and the Vesuvius Challenge have announced a historic breakthrough in the endeavour to decipher text preserved on papyrus scrolls from the ancient site of Herculaneum.

We have this fragment which we believe is from P.Herc 1667, which is one of our scanned scrolls and is roughly 75% segmented. If this fragment is indeed from that scroll, we believe that P.Herc 1667 (Our Scroll 4) is in latin. We have yet to find words from the unrolled scan of this scroll to verify

@scrollprize any known segmentations yet that are in Latin instead of Greek?

Want to be the first person in 2,000 years to read a word in an ancient scroll? Good news! We just released a ton more segmentations from Scroll 3! Head to scrollprize.org or join the Discord discord.gg/uTfNwwecCQ to get started!

@Lorco_me Hey! Shoot an email to press@scrollprize.org. Thank you!

@scrollprize hey guys, who deals with media and speaking requests?

Herculaneum scrolls breakthrough: Scientists digitally "unroll" 2,000-year-old scroll scorched by Mount Vesuvius. cbsn.ws/3Q27rGu

We found some words in Scroll 5, which is from the Bodleian Library collection of Herculaneum papyri. Still work to do to prevent our segmentation algorithm from jumping between adjacent sheets so we can get whole sentences and paragraphs out of this scroll, but this is very promising! Words found include ἀδιάληπτος (‘foolish’), διατροπή (‘disgust’), φοβ (‘fear’), and βίου (‘life’).

In 2023, several tech executives sponsored the “Vesuvius Challenge” competition, offering cash rewards for efforts to decipher the scrolls with machine learning, computer vision and geometry. trib.al/DDmwvue

First glimpse inside burnt scroll after 2,000 years bbc.in/40NJzv2

We recently partnered with @bodleianlibs and @UniofOxford and are thrilled to share some exciting news! Tons of new text, huge segmentations, and updates on the state of ink detection and segmentation: scrollprize.substack.com/p/exciting-new… More to come soon!