OCR vs ADE: Mechanisms Behind the Methods
dev.toยท1dยท
Discuss: DEV
๐Ÿ“„OCR
Scientists May Have Decoded the Mysterious Language of a Lost City
popularmechanics.comยท11h
๐Ÿ—๏ธPaleocryptography
Show HN: Rebuilt Bible search app to run 100% client-side with Transformers.js
biblos.appยท2hยท
Discuss: Hacker News
๐Ÿ“œBinary Philology
Pecia system
rhollick.wordpress.comยท1d
๐Ÿ’งManuscript Watermarks
TALENT: Table VQA via Augmented Language-Enhanced Natural-text Transcription
arxiv.orgยท1d
๐ŸŽ™๏ธWhisper
How JPEG Works: The Subtle Beauty and Art of Imperfection
techspot.comยท1d
๐Ÿ–ผ๏ธJPEG XL
The Dunhuang Culture ๆ•ฆ็…Œๆ–‡ๅŒ– Database
digitalorientalist.comยท10h
๐Ÿ“œText Collation
The artificial complexity of OOXML files (the PPTX case)
blog.documentfoundation.orgยท6hยท
Discuss: Hacker News
๐Ÿ“ŸTerminal Typography
Welcome to LILโ€™s Data.gov Archive Search
lil.law.harvard.eduยท3h
๐Ÿ’พData Preservation
The Aberdeen Bestiary, One of the Great Medieval Illuminated Manuscripts, Now Digitized in High Resolution & Made Available Online
openculture.comยท1d
๐ŸฐManuscript Networks
People rescuing forgotten knowledge trapped on old floppy disks
bbc.comยท10hยท
Discuss: Hacker News
๐Ÿ“ผCassette Archaeology
The Best Ways to Digitize Your Notes
lifehacker.comยท11h
๐Ÿ“„Document Digitization
From Documents to Dialogue: A step-by-step RAG Journey
dev.toยท9hยท
Discuss: DEV
๐Ÿ“ŠMulti-vector RAG
RND1: Simple, Scalable AR-to-Diffusion Conversion
radicalnumerics.aiยท1dยท
Discuss: Hacker News
๐Ÿ’ปLocal LLMs
YouTube gets ~5% CTR lift on Shorts by replacing embedding tables with Semantic IDs
shaped.aiยท23h
๐Ÿ“ŠFeed Optimization
Show HN: Lore Engine โ€“ Turn 10-hour lectures into 2 hours of comprehensive notes
github.comยท1dยท
Discuss: Hacker News
๐Ÿ“„Document Streaming
The Porcelain to Come
stackdiver.comยท57mยท
Discuss: Hacker News
๐Ÿ”ฒCellular Automata
The key to conversational speech recognition
datasciencecentral.comยท1d
๐ŸŽตAudio ML
How the Rise of Tabular Foundation Models Is Reshaping Data Science
towardsdatascience.comยท1d
๐Ÿง Machine Learning
Build-less manga scanlation workflow: vertical OCR, translation, and layout-preserving typesetting in the browser
dev.toยท8hยท
Discuss: DEV
๐Ÿ‡ฏ๐Ÿ‡ตJapanese Computing