O FLISoL, Festival Latino-americano de Instalação de Software Livre ocorre em várias cidades do Brasil e América Latina. Consulte a programação no site flisol.info/FLISOL2026/B... clicando no estado desejado. Discussões sobre segurança. criptografia, Python/R e mais.
#opensource #softwareLivre
Israel ignora cessar-fogo e ataca Líbano matando 300 pessoas. De novo, péssimo dia para este texto, Tábata.
bsky.app/profile/folh...
82% dos israelenses são favoráveis à expulsão de palestinos de suas terras.
www.haaretz.com/israel-news/...
www.cnnbrasil.com.br/internaciona...
Tribunal Penal Internacional emite mandado de prisão contra Netanyahu
... “fome como método de guerra”
De cada 3 mortes de jornalistas no mundo, 2 foram mortos por Israel
cpj.org/special-repo...
Recém-nascidos morrendo de fome em Gaza www.bbc.com/portuguese/a...
Nações Unidas: Uma criança é morta a cada 10 minutos em Gaza
21 Novembro 2023
news.un.org/pt/story/202...
Mas este pelo menos colocava sua marca nos vídeos. Facilitava o trabalho de checagem, mas enganava praticamente na mesma propoção das outras ferramentas, quase ninguém percebia o sora pulando no vídeo. Agora não faltam alternativas.
📢 Muito legal essa ação que o TRE de Mato Grosso fez para combater as fake news e os golpes digitais contra pessoas idosas.
O tribunal levou o projeto “Prosa e Verdade”, do TSE, para tirar dúvidas de mais de 50 idosos em Cuiabá.
Three-panel diagram explaining an audit method for LLM unlearning. Panel A, “LLM Unlearning,” shows a trained language model and a modified “unlearned” model, highlighting that output-based metrics may give a misleading sense of privacy. Panel B, “Our Audit,” illustrates comparing the two models using a Venn diagram to decompose mutual information into unlearned information and residual information. Panel C, “Practical Impact,” shows that residual information can correlate with vulnerability to adversarial attacks, and suggests inference-time privacy protection through abstention.
🤔 𝗗𝗶𝗱 𝘆𝗼𝘂𝗿 𝗟𝗟𝗠 𝗿𝗲𝗮𝗹𝗹𝘆 𝗳𝗼𝗿𝗴𝗲𝘁 𝘆𝗼𝘂𝗿 𝗽𝗿𝗶𝘃𝗮𝘁𝗲 𝗱𝗮𝘁𝗮 - 𝗼𝗿 𝗷𝘂𝘀𝘁 𝗴𝗲𝘁 𝗾𝘂𝗶𝗲𝘁𝗲𝗿 𝗮𝗯𝗼𝘂𝘁 𝗶𝘁?
🧠 Machine unlearning aims to remove specific training data (e.g., private info) without full retraining. But does it actually 𝗿𝗲𝗺𝗼𝘃𝗲 𝘁𝗵𝗲 𝗸𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲 𝗳𝗿𝗼𝗺 𝘁𝗵𝗲 𝗺𝗼𝗱𝗲𝗹?
Accessible for all here: doi.org/10.31235/osf...
I made a map of 3.4 million Bluesky users - see if you can find yourself!
bluesky-map.theo.io
I've seen some similar projects, but IMO this seems to better capture some of the fine-grained detail
🚀 dashboardr is out!
An #rstats package for building interactive dashboards with tidyverse-style syntax.
Launched with a fun hackathon at @ascor.bsky.social (🍕 included).
📦: favstats.github.io/dashboardr/
Big thanks to the dashboardr team, Digicomlab, and @vivifabrien.bsky.social for the logo 🎨
Together with Lisa Ginsborg we are happy to share that our edited volume on "Disinformation: A Multidisciplinary Analysis" is now published!
A long journey which involved many esteemed authors, with whom it was a great pleasure to work and to which we are grateful
link.springer.com/book/10.1007...
✨ We’re excited to announce the Spring 2026 IAS Seminar Series, featuring a stellar lineup of speakers and thought-provoking talks. Open to all! #AcademicSky
Meet Netha, the Wikipedia editor who quietly debunks medical myths for millions #Wikipedia25
@carapanarana.bsky.social @bicicreta.bsky.social
🇫🇷 We are hiring 🇫🇷
Assistant or Associate Professor Position in Computational Sociology @crestsociology.bsky.social @ipparis.bsky.social
Details here (please RT)
www.shorturl.at/E57le
A política externa de Trump não opera em eixos tradicionais –aliados versus adversários, democracias versus autocracias, competição estratégica versus cooperação. Opera em um cálculo mais simples: Você consegue revidar com força suficiente para machucá-lo? Se a resposta for não, e você tem algo que ele quer, você é um alvo. Se for sim, ele fará um acordo.
www1.folha.uol.com.br/colunas/ian-...
Me pareceu acertado trocar hemisfério de influência por lei da selva.
semantic network of Trump's speeches today about Maduro's arrest.
Example of Trump's phrases about oil: "Well, in terms of China and Russia, yeah, uh, well, Russia, when we get things straightened out, but in terms of other countries that want oil, uh, we're in the oil business, we're gonna sell it to them."
I scraped a transcript of Trump, Rubio, etc. speech today about Maduro's arrest. It is freely available at this link. One curiosity: Trump’s speeches about Maduro’s arrest featured oil in 17 phrases and democracy in none.
#rstats
docs.google.com/spreadsheets...
The great advantage of having multiple bathrooms is the variety of windows as entry points, as there is no door to enter the house.
Para quem se interessa por #HumanidadesDigitais, está dias 26 e 27, o 3rd International Conference on Data & Digital Humanities (DDHUM2025), gratuito para assistir online, presencial em Braga, Portugal. A maioria das falas tem sido em português e algumas em inglês.
ceh.elach.uminho.pt/ddhum
new paper by Sean Westwood:
With current technology, it is impossible to tell whether survey respondents are real or bots. Among other things, makes it easy for bad actors to manipulate outcomes. No good news here for the future of online-based survey research
The merchants of doubt 💰
www.theguardian.com/environment/...
The documents, which include copies of the actual cheques Exxon sent, consist of internal documents and years of correspondence between... Atlas Network, a US-based coalition of more than 500 free-market thinktanks...
#disinformation
A screenshot from the Gab AI page advertising its AI tools as promoting Christian values and free thinking: “The One AI They Can’t Control.” They provide an example of responses from different AI models to the question “who is the biggest perpetrator of political violence in America,” with ChatGPT and Gemini answering “right-wing” and Gab’s Arya answering “left-wing.
I was analyzing data from X/Twitter from 2023 concerning #disinformation and #conspiracytheories and discovered
the link to the Gab's AI
"The One AI They Can't Control", "Our exclusive AI model trained to uphold Christian values and traditional principles in every interaction."
If you frequently run #LLM models locally in Linux through the command line using #ollama, your life can be a little bit easier with ollama-run. This simple shellscript frees you from typing the name of the model and the possibility of misspelling it.
github.com/SoaresAlisso...
Abstract
New paper out with @dasalgon.bsky.social: “Far-Right Agenda Setting: How the Far Right influences the Political Mainstream” doi.org/10.1017/S1475676525100066 #openaccess in @ejprjournal.bsky.social🧵
"U.S. President Donald Trump was the dominant source and
spreader of #disinformation in our content analysis.... There is a correlation between the rise of Trumpism on one hand and the loss of trust in journalists as ‘watchdogs’/‘truth seekers’ on the other, "
www.icfj.org/sites/defaul...
Important point that I forgot to mention. The allegation that universities are producing drugs came from the former right-wing Brazilian education minister.
exame.com/brasil/minis...
At least in Brazil, the far right began by criticizing the humanities and quickly moved on to claims that “universities are woke” and “public universities are drug factories,” cutting funding for all higher education. The humanities are the first step, then they go to biology, health sciences...
The "validate, validate, validate" (GRIMMER, 2014) principle of Text Analysis/NLP never gets old.
If you feel uneasy using LLMs for data annotation, you are right (if not, you should). It offers new chances for research that is difficult with traditional #NLP/#textasdata methods, but the risk of false conclusions is high!
Experiment + *evidence-based* mitigation strategies in this preprint 👇
Transformers Models Timeline
Interactive timeline to explore models supported by the Hugging Face Transformers library
#transformers #LLM #bert #IA
huggingface.co/spaces/yonig...
