Would love your feedback on this: AI Systems = Code (model/algorithm) + Data. Most academic benchmarks/competitions hold the Data fixed, and let teams work on the Code. Thinking of organizing something where we hold the Code fixed, and ask teams to work on the Data. (1/2)

Tremendous thanks to @simon_wiles for his great kindness -- incorporating this long desired toolset from @johnrladd into Palladio -- and to @quinnanya for her care and attention to keeping the work of Giorgio Caviglia and Ethan Jewett available for data exploration.

Calling all librarians, archivists, curators in NYC.

It can feel like there's little we can do #UkraineUnderAttaсk but we can do a little. Getting a warm glow working on Saving Ukrainian Cultural Heritage Online sucho.org, seeing so many #DH friends @ekansa Scott B. Weingart Kim Martin. Come and join us. It's for everyone.

Many thanks to @ETreharne, ummmm, Mark Algee-Hewitt, and @jeffssteward

Scott B. Weingart alerted me that our book (co-written w Nicole Coleman and Sebastian Ahnert) The Network Turn has inspired a journal special issue CFP! So lovely it’s started a conversation - this was our hope! avldigital.de/vernetzen/deta…

We will look at the morbidity and mortality impact of this disgraceful decision. And those who made will have to be held accountable. #shame #WomensRights

Excited to announce #BigLAM, a BigScience Research Workshop x Hugging Face Hackathon focused on making library, archive, and museum datasets related to machine learning more accessible and discoverable github.com/bigscience-wor…

I've been playing with the training dataset behind the extremely cool new BLOOM model from BigScience Research Workshop and Hugging Face. Here's a sample of 10 million chunks from the English-language corpus, about 1.25% (!!) of the total. Encoded with `all-distilroberta-v1`, then UMAP to 2d.

Scott B. Weingart Nicole Coleman Sebastian Ahnert I spy our little book.

Excited to share a new two-part Programming Historian lesson, "Computer Vision for the Humanities: An Introduction to Deep Learning for Image Classification." programminghistorian.org/en/lessons/com…

These points by Lynn Cherny contain important guidance for successful implementation

🎶 Five hundred pages of copyediting on the wall... 500 pages of copyediting... take one down, pass it around, 499 pages of copyediting 🎶 -- and it's done! The Dangerous Art of Text Mining is coming out soon from Cambridge University Press ! 🍾

Last week I presented this paper as a (non-archival) poster at ACM EAAMO (Handle Retired, Visit @EAAMO_ORG) conference: A Human Rights-Based Approach to Responsible AI arxiv.org/abs/2210.02667 Joint work w/ MMitchell @timnitGebru (@dair-community.social/bsky.social) Iason Gabriel, started back in 2020! Excited to have it be out there, finally!

We've just released from Nomic AI a new map for exploring 6+ million AI-generated images and the prompts used to created them, collected by KREA AI. atlas.nomic.ai/map/809ef16a-5… The most exciting thing for me here is how this changes full-text search: here's why (thread).

The Arcade Interventions editor, Charlotte Lindemann has published a series of interviews about DH with Stanford faculty and staff including Giovanna Ceserani, Mark Algee-Hewitt, Alice Staveley, Alix Keener and myself. shc.stanford.edu/arcade/interve… My msg: Why funding for humanities is critical.