Étiquette : google (Page 15 of 35)

«Les prétentions à la suzeraineté de ces plates-formes, pour reprendre la formule de Alain Supiot (2) dans Gouverner par les nombres, doit être combattue par la souveraineté des États (comme le fait le RGPD) pour contraindre ces plates-formes à remettre dans le domaine public toutes les données et traces qui sont à leur disposition, données produites par le public et qui méritent d’être reprises, traitées et contrôlées par des organismes scientifiques publics avant d’être mises à disposition des chercheurs dans des conditions précises comme c’est le cas pour les recensements (et donc avec des limites). Les principes de la science ouverte risquent fort de rester des vœux pieux si on ne s’attaque pas dès maintenant à cette prédation des données par l’oligopole de l’attention que constituent les GAFAMT» – Dominique Boullier.

Source : Facebook et la recherche : le « quasi État » | InternetActu.net

Decentralize the Internet

«Finding the killer apps of the decentralized internet will take more time, people, and money than have been thrown at the problem so far. Pakman says that societal attitudes to power and big tech companies appear to be in the right place to deliver them. ‘There’s massive distrust in centralized everything,’ he says. ‘We don’t trust the government, don’t go to church or synagogue, don’t trust banks and now we no longer trust tech companies.’»

Source : The Decentralized Internet Is Here, With Some Glitches | WIRED

«With that, Hangouts Chat is now ready to go up against Slack (6 million users as of last September), Microsoft Teams, Facebook Workplace, Atlassian’s Stride, Flock, Zoho Cliq, and Keybase’s free Teams. Google grew its G Suite customer base by a third over the past year to 4 million businesses, so it could well have a fighting chance in this battle».

Source : Google’s Hangouts Chat for teams is now globally available

«We present a method to create universal, robust, targeted adversarial image patches in the real world. The patches are universal because they can be used to attack any scene, robust because they work under a wide variety of transformations, and targeted because they can cause a classifier to output any target class. These adversarial patches can be printed, added to any scene, photographed, and presented to image classifiers; even when the patches are small, they cause the classifiers to ignore the other items in the scene and report a chosen target class».

Source : Adversarial Patch – Research Article | DeepAI

«Goodfellow’s friends were just as adamant that this method wouldn’t work, either. So when he got home that night, he built the thing. « I went home still a little bit drunk. And my girlfriend had already gone to sleep. And I was sitting there thinking: ‘My friends at the bar are wrong!' » he remembers. « I stayed up and coded GANs on my laptop. » The way he tells it, the code worked on the first try. « That was really, really lucky, » he says, « because if it hadn’t of worked, I might have given up on the idea. »»

Source : Google’s Dueling Neural Networks Spar to Get Smarter, No Humans Required | WIRED

« Older posts Newer posts »

© 2026 no-Flux

Theme by Anders NorenUp ↑