Full transcription of Geneva 146

We have just uploaded a full transcription of the entire 590 pages of the Geneva 146 manuscript from e-Codices based on the collated user-transcriptions to the brandnew eScriptorium website. We have simply used collatEx to align the user transcriptions on a word level and chosen the most frequent proposal. We will now analyse the result in order to improve the crowdsourcing workflow and validate the transcription. Many thanks to the more than 200 transcribers for their awesome work and to the colleagues from the Swiss e-codices project for the gorgeous pictures!! All this will be presented at DH2019 in Utrecht. See you there!

Tikkoun Sofrim at SBL 2019

We have been notified that proposal “Automatic Transcriptions of Medieval Hebrew Manuscripts and Crowdsourcing their Corrections” has been accepted for presentation at the SBL by the 2019 Annual Meeting program unit Digital Humanities in Biblical, Early Jewish, and Christian Studies. Looking forward to San Diego! We will keep you posted.

upcoming talk at Princeton University

Tikkoun Sofrim, Sofer Mahir, eRabbinica and eScriptorium will be presented at a lecture entitled

Artificial Intelligence, Citizen Science and Medieval Hebrew Manuscripts

at Princeton University on Friday April 5, 2019 at noon in room 203 in the Scheide Caldwell house. If you intend to come please inform Baru Saul from the Judaic Studies programme

upcoming presentation at U Maryland MITH

Tuesday, April 2, 2019 – 12:30pm to 1:45pm
MITH, 0301 Hornbake Library
Daniel Stoekl Ben Ezra
 
Recording of the whole talk available on youtube
For Tikkoun Sofrim check out 35:40-50:00.

30.000 transcribed lines

Today, March 19, we passed the threshold of 30.000 transcribed lines. We accumulate 5 transcriptions for each line. So far, more than 6.000 lines, more than half of the manuscript, have already been transcribed by at least one user in the 35 days this project has been running. Amazing! יישר כוח!

We are hurrying up the production of the new interface that will allow multiple manuscripts to be treated at once and a more extensive user feedback is coming, too.

Inauguration of Crowdsourcing Website

We are pleased to announce the launch of a project that combines automatic transcription of Medieval Hebrew manuscripts with the wisdom of the general public. After much work, our website opens today. Please check it out!

As of today and for 34 days of study in Deuteronomy in the Israeli 929 project, the public is invited to assist in the correction of the automatic transcriptions of manuscripts provided by kraken.

The use of the site is quite simple. In the center of the page is the workspace, with an image of the manuscript line and its automatic transcription to be corrected by you.
There are alsp some buttons with which you can mark special characters or special effects in the text.

On the right is a help area. Here you can see e.g. the entire page from which the line was taken and use it for displaying another part of the page. You can check hard letters against others, or you can read detailed explanations of all the special phenomena – Special characters (such as the name of God) appear in it – and how to correct them.

Your feedback is very important to us! Comment below or write to us at tikkoun.sofrim@gmail.com or consult the Eliahu Lab’s Facebook page, where he has a special post for consultations and training.

More updates – Coming soon …

 

Tikkoun Sofrim

This blog is dedicated to the PHC Israeli-French project Tikkoun Sofrim on Machine learning and gamification applied to crowdsourcing corrections of automatic transcriptions of Medieval Hebrew manuscripts funded until december 2019.
Project partners are Haifa University and the National Library of Israel as well as the Ecole pratique des hautes études at PSL University in Paris. PIs are Moshe Lavee and Tsvi Kuflik on the Israeli side and Daniel Stoekl Ben Ezra on the French side.
The blog will update the community about the project’s progress, news and events.