Julkaistiin tänä keväänä kaikille avoin MOOC-kurssi, joka käsittelee digitalisoituvaa yhteiskuntaa ihmistieteiden näkökulmista. Tämä on yhtä tärkeää teknologiaosaamista kuin tekoäly- ja koodauskurssit!
rajapinta.co/2024/05/21/kaik…
O retweeted
I'm happy to announce a new release of AntConc (ver. 4.3.0) that integrates traditional corpus methods with Large Language Models (LLMs) via a new 'ChatAI' tool. I've kept it as a beta version for now, so I can get everybody's feedback. I hope you like it! laurenceanthony.net/software…
O retweeted
Announcing surya - a multilingual text line detection model for documents. It gives you accurate line-level bboxes and column breaks.
Find it here - github.com/VikParuchuri/sury… .
O retweeted
Everyone building RAG uses dense embedding retrieval, but simply doing cosine distance doesn’t always capture fine-grained similarity.
That’s why SOTA retrieval like ColBERT models are so important; these new architectures are fast but more powerful than pure dense retrieval.
@bclavie’s new RAGatouille project makes it super easy to use ColBERT out of the box or fine-tune it over your data. In turn, we’ve made it super easy to plug it into your @llama_index pipeline! Check out our brand-new LlamaPack here:
llamahub.ai/l/llama_packs-ra…
If you’re interested in using ColBERT as an e2e index in @llama_index, we also have a `ColbertIndex` (s/o @Haotianzh): github.com/run-llama/llama_i…
Img credits to the orig ColBERT paper by @lateinteraction et al.
O retweeted
We just published version 1.2 of HPLT datasets. What's new?
- we fixed a bug in monolingual dedup, please redownload! 🛠️
- we filtered out very ugly monolingual documents🤮
- we anonymised the bilingual datasets🕵️♀️
hplt-project.org/datasets/v1…
O retweeted
Free MIRRORSHADES! Weirdly this epic cyberpunk antho doesn't exist as an ebook. So we turned it into a free online webpage. Happy reading, and may ye wax gnarly and punk and dirty and ecstatic and intricate and all that good shit. Happy cybermonday. rudyrucker.com/mirrorshades/
O retweeted
🥰16 Network visualization tools. In this 2nd version of my poster, I updated some networks, added more tools, and wrote a Medium Story.
📎 medium.com/@vespinozag/16-ne…
#NetworkScience #AcademicTwitter #DataVisualization #postdoc #MachineLearning #DataScience #Bioinformatics #PhD
O retweeted
finally published our latest research on text embeddings!
TLDR: Vector databases are NOT safe. 😳 Text embeddings can be inverted. We can do this exactly for sentence-length inputs and get very close with paragraphs...
Heinäkuussa ilmeistyi @kultutlehti teemanumero "Mittaamisen mitta" jossa katsaukseni otsikolla "Sosiaalisen median tutkija ja houkuttelevat numerot". Somen numerodataa kannattaa tutkimuksessa - ja muutenkin - käyttää kriittisesti harkiten.
journal.fi/kulttuurintutkimu…
On typeriä päätöksiä, on typerämpiä päätöksiä ja on tämä päätös. Yksikään sosiaalisen median alusta ei yritä RAJOITTAA käyttäjien alustalla viettämää aikaa. Mutta uudet (väliaikaiset) rajoitukset ovat:
To address extreme levels of data scraping & system manipulation, we’ve applied the following temporary limits:
- Verified accounts are limited to reading 6000 posts/day
- Unverified accounts to 600 posts/day
- New unverified accounts to 300/day
1
163





















