Show more

@davidrevoy For those (like me) who are out of the loop: arstechnica.com/information-te.

TL;DR: DeviantArt created an art generator called DreamUp based on Stable Diffusion, and art uploaded by the site’s users will be used for training unless they opt out.

Great news #Fediverse it is Official. The #EU #EuropeanCommission just launched two servers 🚀

Welcome to #EUVoice mastodon and #EUVideo peertube 🎉

So say hi to @EC_Commission@social.network.europa.eu 👋

Love #FOSS with @EC_OSPO 💘

Stay in the loop with @EC_DIGIT

Help make things fundamentally right with @FRA ☀️

Regain balance with @Curia ⚖️

Protect that data with @EDPS 🔐

Let's go global with @CDT 🌐

Get us heard at @ombudsman 📣

Watch and boost from: tube.network.europa.eu/videos/

And explore: social.network.europa.eu/explo

(1/6)

Summarizing this excellent systematic survey of 507 papers on the state of #KnowledgeGraph in #NLP since the first Internet-age #KG was announced 10 years ago:

Schneider et al. 2022. “A Decade of Knowledge Graphs in Natural Language Processing: A Survey.” arxiv.org/abs/2210.00105

1. Tasks surveyed are broadly divided into Knowledge Acquistion and Knowledge Application.

#PaperThread

The first international workshop on Legal Information Retrieval will be held at ECIR 2023, in Dublin, on April 2nd, 2023 #ECIR2023 #LegalIR

The call for papers (actually extended abstracts) is now available on tmr.liacs.nl/legalIR/

Deadline: January 20th, 2023

We test models from @huggingface hub
rank them efficiently (linear probing on one task)

The best ones, we finetune on 36 different datasets and share with you here:
ibm.github.io/model-recycling/

Next time you finetune, just pick the best one

Why use a worse model?

#NLProc #nlp #MachineLearning #ml
Fixed link...

Show thread

RT @MinqiJiang@twitter.com

SAMPLR allows adaptive curricula to focus more on difficult scenarios (e.g. “driving on ice”), without the final policies *overestimating* their probability. In other words, it robustifies agents to difficult settings, without making them overly conservative or optimistic.

🐦🔗: twitter.com/MinqiJiang/status/

I don’t train from scratch, I use RoBERTa🧐
Wait…
Why not cross-encoder/stsb-roberta?facebook/muppet-roberta?

We automatically identify the best models on 🤗(periodically)

Just pick the best one
and finetune on your task

ibm.github.io/model-recycling/

#NLProc #MachineLearning #finetuning #intertraining #huggingFace #SoTA

SigmoidS has exceeded 2k registered users!

Keep the trend going by letting your friends/colleagues know about what a cool place this is becoming for AI folks :) We'll keep scaling up the server compute as needed as more users join.

Since some have asked: you can support us monetarily by subscribing on Substack - thegradientpub.substack.com/

We are considering setting up a separate Patreon just for Sigmoid Social, but for now feel free to help us out that way if you'd like :D

I have been researching and building foundations to link text to knowledge, constructing like , and bringing together unstructured and structured knowledge to power a variety of applications.

I am now working on pushing the boundaries of as CTO for the AI unit at SAP.

I am very much hoping to keep track of the latest in and here, and will occasionally contribute my own view and works :)

#Introduction I am hoping to join a welcoming and diverse community here. What I liked about Twitter was the convos about ML, the new papers, ideas, and feeling connected to other researchers across the globe. Looking forward to rebuilding this on an open platform without the volatile billionaire bit.

I’m a professor of CS and will occasionally post something about geometric (graph) deep learning, attempts to bridge discrete and continuous learning, and applications in the sciences.

Time for an #introduction:

I am a researcher at Google Brain 🧠 in Amsterdam 🚲, working primarily on structured approaches to ML/AI: graph representation learning / GNNs, structured / geometric scene representations & world models, object-centric learning & compositionality.

You probably won't see me toot/tweet very often, but when I do, it will mostly be about research that excites me and the occasional toot about car-free / livable cities.

Thank you, @thegradient for setting up this server!

Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.