#ff2024 — Public Fediverse posts

Am 06.12. 🗓️findet unser 3. #OpenData 📂Fachforum #FF2024 statt! ‼️
Dieses Mal mit drei Vorträgen zum Thema
#Datenqualität & #Standardisierung !
Wir freuen uns auf spannende Vorträge und einen regen Austausch!🙌
Eine Nachlese findet Ihr im Anschluss auf:👇
https://www.opendata.bund.de

@hugh
That was a fantastic write-up! Thank you!

#AI #Ethics #SlowAI #FF2024

Great #FF2024 reflection post by @hugh. "Helping users take a 'sense-making' journey."

https://ausglam.space/@hugh/113501858032432797

This and #FF2024 in general make me think we need updated studies on information seeking in different types of GLAM collections. And that all modern searches need a 'semantic search off' toggle to enable exact matches again

https://ausglam.space/@hugh/113501858032432797

What a delight it was to present at the #ff2024 #ff24 Fantastic Futures #conference last week, held at the #NFSA, celebrating all things #AI and #ML in the #GLAM sector.

See what the #ANU School of #Cybernetics got up to at the conference:

https://cybernetics.anu.edu.au/news/2024/10/18/fantastic-futures?utm_source=mastodon&utm_medium=social

A great example of learning on the job from Asa - 'Making cursive comfortable: Transcribing 19th Century archives using AI models' https://prov.vic.gov.au/about-us/our-blog/making-cursive-comfortable as presented at #FF2024 #AI4LAM

As if prompted by #FF2024, RNZ ran this article featuring Peter Lucas-Jones and the great 92% accuracy Te Reo Māori text-to-speech work at Te Hiku Media.

"It is important to recognise that big tech is not going to save te reo Māori."

https://www.rnz.co.nz/news/alert-nat/531303/why-google-maps-can-t-pronounce-maori-place-names

#AI4LAM

#FF2024 ended with a literal bang in the form of Triptych, EDM+AV Mechanical Synaesthesia, by Robin Fox. It was loud, intense, immersive, and what could only be described as an unforgettable "trip" 🔥🔥🤩💫💥. Nothing I've ever experienced before. A reimagining. In many ways, that was one of the main threads at #FF2024 - deliberate and cautious approach, ask questions, be more conscientious, do #SlowAI if possible. Certainly a reimagining of how AI projects have been done post-Kraken 🐙 👹🧌.

#AI4LAM

Homeward bound from a great #FF2024 #AI4LAM on the smallest plane I have been on for a while!

Just like @mia, this talk by Barbara Shubinski, Rockerfeller Archive Center, drew me in, and I couldn't summarise. It would be great as a full talk. Maybe at #FF2025
https://hcommons.social/@mia/113326897490041469

#FF2024 #AI4LAM

#AI4ALAM folk, you outdid yourselves at #FF2024! Your optimism, your cynicism, your questions about how we get the AI / ML we deserve - and most importantly your support for each other. Please keep sharing and learning together! https://ai4lam.org

@nfitzger announces that #ff2025 will be in London 3-5th December 2025 at the #BritishLibrary
#AI4LAM #ff2024

#SaveTheDate 💾 #FF2025 at the British Library! 3-5 December 2025.🔥🤩🤩🔥

#FF2024 #AI4LAM

Karen Thompson – Applying AI to accelerate data transcription from digital objects
University of Melbourne

Hespi. Sheet component model Built using Object Detection Model YOLOv8 (built in-house). Model freely available. Labels freely available.

✉️ Contact Karen, use the model, train the model, use it in your institution. Details in 2nd image.

#FF2024 #AI4LAM

Liked #FF2024? Come to London for #FF2025!

3 - 5 December 2025

https://ai4lam.org/ #AI4LAM

We hope you enjoyed #FF2024! Save the date for #FF2025 Fantastic Futures 2025 AI4LAM at the British Library, London 3-5 December 2025 https://ai4lam.org

Kara Kennedy, AI Literacy Institute – AI Literacy for Librarians in Low-Resource, Multicultural Communities

https://ailiteracy.institute/

Two choices, don't use AI (like early days of Wikipedia) or learn how to use AI wisely
1. Engage AI champion
2. Leverage existing content - LinkedIn learning. Lots of free resources.
3. Create accessible resources

Librarians, GLAM, we can lead AI literacy, help prepare community for AI

#LeaveNoOneBehind

#FF2024 #AI4LAM

Barbara Shubinski, Rockefeller Archive Center – Gender in a (Supposedly) Non-Gendered Historical Space: Generative AI, Sentiment Analysis, and Graph Databases in the Rockefeller Foundation Collections of the 1930s - impossible to summarise but compelling talk! #FF2024 #AI4LAM

End up, created our own - Whisper Scribe. https://www.github.com/gsu-library/whisper-scribe (Rachel invited us to check it out)

Future development: partitions, dynamic user permissions, locking records when editing, investigate syncing existing transcript, project profiles.

#FF2024 #AI4LAM

Rachel Senese Myers – Adopting Whisper: Creating a front end optimized for processing needs
Georgia State University

Oral history. Ohms. Not searchable.

Experiment: faster Whisper. Took learnings from #FF2023. Blew Adobe Premiere Pro out of the water. IT said, you can't do this - command line. Need UI.

Whisper Web UI.
No database. Browser crash, data gone.

Whishper. Love database. No batch upload. No diarization. No batch edits.

#FF2024 #AI4LAM

#FF2024 Rachel Senese Myers - Georgia State web ui for Whisper - go check their github! #AI4LAM

Lightning talks (5 mins each).

Asa Letourneau – C19th Handwriting and Model Making with Transkribus
Public Record Office Victoria, State Archives Victoria

Rapid learning and experimenting. Started July 2023. Tried Trankribus. Not successful. Created 3 private model, CER 5%. Tested public model Text Titan, 2.75% CER.

Now Trankribus beta site. Recognise tabular record. Hoping to deploy IIIF Glycerine for crowdsourcing editing of transcription.

#FF2024 #AI4LAM

Asa Letourneau talking #OCR and handwriting recognition with #Transkribus at Public Records Victoria. From no knowledge of #AI to a working service in 12 months. Reflecting on learnings about #DataSovereignty.

#ff2024 #AI4LAM

Asa Letourneau on Transkribus at Public Record Office Victoria - will start crowdsourcing corrections. From newbie in July 2023 to this... #FF2024 #AI4LAM

Session 6 Institutional Adoption, talks by Lizabeth Johnson (LANL), Barbara Shubinski (RC), Kara Kennedy, Karen Thompson (UMELB), and Neil Fitzgerald (BL), hosted by Jess Herrington (ANU) #ff2024 #ai4lam

https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.ebrmxvadj9i8
https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.pbldge26q2x3
https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.2mz94hhg68w5
https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.9fekuu25s6pz

Human to judge generated similarity. 900 judgements by 15 NLA staff. Head to head comparison. Individually, CLIP highest score. Combination of model considered best 66% of the time against other variants. Open source models getting better.

#FF2024 #AI4LAM

Nice comparison tool from NLA to see which models did better at text search results and image similarity searches #FF2024

🗣️ Session 6 AI and Institutional Adoption, talks by Laura McGuinness (LANL), Kate Follington & Asa Letourneau (PROV), Rachel Senese Myers (GSU), hosted by Jess Herrington (ANU) #ff2024 #ai4lam

https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.9h1gqz6m5cz6
https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.hsn26u9bzgd9
https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.r2llfyk7dhx2

Evaluated OpenAI CLIP (Jan 2021), OpenAI gpt-4-vision-preview (proprietary, Jan 2024) and Microsoft Phi-3.5 Vision (open source, Aug 2024). No data retention models.

Nomic, open source system, generate embeddings. Store in Solr. Incoming text query to embeddings using Nomic. Then KNN query to Solr.

CLIP uses its own embeddings.

Image: Fixed query set. Fixed query types.

#FF2024 #AI4LAM

Miguel: Project looking for friends (partners) and funding #FF24 😜😁

#AI4LAM #FF2024

#FF2024 Morgan Strong (QAGOMA) 
ArtSEEker app can recognise collection artworks, find five predominant colours then search for other works with the same balance of colours - IMO that's such a generous, chilled way to search art collections
#FF24 #AI4LAM

Kirsten: On #SlowAI and "Just chill". Cultural understanding of time and space and slowing down. Fear of inertia.

Lauren: Yarning as methodology. "Isn't this just a focus group?" Yes, it does sound like it. But it's much much more. It encapsulates relationality btw research participant and researchers, conflict of interest. Recognising of prior relationship. Adds another layer or accountability. It needs to be Indigenous-led, can't just tack something on.

#FF2024 #AI4LAM

#FF2024 Jon Dunn reminds us all that you can join informal working groups on speech-to-text and more at https://ai4lam.org - anyone is welcome to join the Slack, attend online calls with demos etc #AI4LAM #FF24

James Smithies & Karaitiana Taiuru – Large Language Models and Transnational Research: Introducing the AI as Infrastructure (AIINFRA) Project.

Goal to have an evaluation framework for AI as a research tool. #SlowAI. Incorporate indigenous knowledge protocols. Sources: Hansard 1901 from UK, Aus and NZ.

Aotearoa - Data sovereignty principles created by elders and community, not by academics.

Abstract at https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.ytx09syhqiuq #AIINFRA #AI4LAM #FF2024

🗣️ Session 2 AI and Conversational Archive, talks by Peter Leonard (StanfordU), Mia Ridge (BL), Joshua Ng (ANZ) & Peter Broadwell (StanfordU), hosted by Shahab Qamar (NFSA) #ff2024 #ai4lam #glamr #dh

https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.7s7i6tjputwx
https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.l195pmhmy9kb
https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.h3tolnbck1k8
https://sites.google.com/view/ai4lam/news/fantastic-futures-2024-papers#h.yjgpb2jpyg8l

Get in touch with Grant and others at the NFSA about their Bowerbird transcription project at https://nfsa.jotform.com/form/242831153056855 #FF24 #FF2024 #AI4LAM #WhisperAI #MuseTech

##F24 @KathyReid - hearing a lot of conversations about the need for Australian GLAMs to adopt AI but fewer conversations about where that AI comes from - colonisation by Silicon Valley. Cultural institutions could help develop national capabilities in AI. She wants a Whisper that works for all Australian speakers #FF2024 #AI4LAM

##F24 @KathyReid - hearing a lot of conversations about the need for Australian GLAMs to adopt AI but fewer conversations about where that AI comes from - colonisation by Silicon Valley. Cultural institutions could help develop national capabilities in AI. She wants a Whisper that works for all Australian speakers #FF2024 #AI4LAM

##F24 @KathyReid - hearing a lot of conversations about the need for Australian GLAMs to adopt AI but fewer conversations about where that AI comes from - colonisation by Silicon Valley. Cultural institutions could help develop national capabilities in AI. She wants a Whisper that works for all Australian speakers #FF2024 #AI4LAM

##F24 @KathyReid - hearing a lot of conversations about the need for Australian GLAMs to adopt AI but fewer conversations about where that AI comes from - colonisation by Silicon Valley. Cultural institutions could help develop national capabilities in AI. She wants a Whisper that works for all Australian speakers #FF2024 #AI4LAM

The next meeting of the Research Data Alliance's Collections As Data Interest Group will be held on 18 September, 3.00-4.00pm AEST. It'll include presentations from: Daniel Wilksch (Public Record Office Victoria) who’ll describe the process of developing PROV’s collection API; and @ingridbmason who’ll introduce #AI4LAM and preview the Fantastic Futures conference in October. More details, registration link, and agenda here: https://www.rd-alliance.org/groups/collections-as-data-ig/forum/topic/next-rda-collections-as-data-ig-meeting-18-september/ All welcome! #GLAM #openData #CollectionsAsData #ff2024

long toot about unauthorised web scraping of works of art by ‘AI’ (🙄) / LLM crawlers & possible countermeasures. boosts appreciated! 🙏

(thanks to @KathyReid for some new (to me) hashtags to reach out on, doing a new toot rather than hijacking hers 😜)

so, atm I’m focused on possible countermeasures to #WebScraping & absorption of works of art & commentary around them into #img2img & #text2img #AI (🙄) / #LLM models - but it feels like the only viable long-term controls aren’t technical but rather regulatory & legislative, which tend to have long lead-times, limited consequences for violators (often small fines or a stern warning 🙄), & multi-jurisdiction issues & loopholes 😕

when it comes to technical controls… robots.txt is basically relying on the honour system (fails when crawlers have no honour), some interesting approaches with poisoning tools like glaze & nightshade (only useful before the scrapers hit, subject to counter-countermeasures, & potential legal issues for those obliged to care about such things)… 😢

some CDNs seem to do a decent job of #fingerprinting crawlers based on HTTP headers & subtleties in packets, etc., & dropping their traffic just for bandwidth concerns, but it feels like we’re in an arms race & have one arm tied behind our backs while our adversaries just yolo their way around such trivial concerns as copyright, licensing, artists’ moral rights, & seem to actually think that ‘but if we had to pay for the training content, our business model just wouldn’t work!’ (yeah? and?) is a thing that anyone should care about… 🤔

if anyone following these hashtags has advice about possible defences or anything else I haven’t seen yet, I’m all 👂👂👂👂👂

#FF2024
#AI4LAM
#GLAM