Auteur/autrice : noflux (Page 50 of 625)

Facebook AI moderator confused videos of mass shootings and car washes | Ars Technica

13 novembre 2021 /

A frowning man in a business suit.

“Facebook’s internal documents reveal just how far its AI moderation tools are from identifying what human moderators were easily catching. Cockfights, for example, were mistakenly flagged by the AI as a car crash. “These are clearly cockfighting videos,” the report said. In another instance, videos livestreamed by perpetrators of mass shootings were labeled by AI tools as paintball games or a trip through a carwash. If the situation sounds grim in the US or among English-speaking countries, it appears far worse elsewhere. In Afghanistan, for example, the company said in reports that it lacks a dictionary of slurs in the country’s various languages. As a result, Facebook estimates that it identified just 0.23 percent of hate speech posted on the platform in Afghanistan.”

Source : Facebook AI moderator confused videos of mass shootings and car washes | Ars Technica

12 novembre 2021 /

“At YouTube, we strive to be a place where creators of all sizes and backgrounds can find and share their voice. To ensure that YouTube promotes respectful interactions between viewers and creators, we introduced several features and policies to improve their experience. And earlier this year, we experimented with the dislike button to see whether or not changes could help better protect our creators from harassment, and reduce dislike attacks — where people work to drive up the number of dislikes on a creator’s videos.
As part of this experiment, viewers could still see and use the dislike button. But because the count was not visible to them, we found that they were less likely to target a video’s dislike button to drive up the count. In short, our experiment data showed a reduction in dislike attacking behavior.
We also heard directly from smaller creators and those just getting started that they are unfairly targeted by this behavior — and our experiment confirmed that this does occur at a higher proportion on smaller channels.
Based on what we learned, we’re making the dislike counts private across YouTube, but the dislike button is not going away. This change will start gradually rolling out today.”

Source : An update to dislikes on YouTube

Materialism — Studio Drift

5 novembre 2021 /

I-phone 4 from 2010 - work 2019

“Materialism confronts the viewer on a very elementary level with the things we surround ourselves with and the materials that comprise them. The work calls for contemplation on how people deal with the raw materials at their disposal.Everyday products such as a vacuum cleaner, Volkswagen Beetle, pencil, or PET bottles, have been reduced to the exact quantity of the specific raw materials from which they are made, shown in the form of rectangular blocks. ”

Source : Materialism — Studio Drift

Squid Game Cryptocurrency Scammers Make Off With $2.1 Million

5 novembre 2021 /

The price of SQUID cryptocurrency plunging to $

“If you’re going to buy cryptocurrency, the most important thing to look out for isn’t necessarily the price. First, figure out whether you can sell the coin after you bought it. If you can’t sell—like the people who invested in SQUID discovered they can’t—it doesn’t matter how high the price goes, just as it doesn’t matter how large of a number someone writes on Monopoly bills. Cryptocurrency is only worth what someone else is willing to pay for it. And if the rules say you can’t even sell, you’ve just been scammed. Sorry, folks.”

Source : Squid Game Cryptocurrency Scammers Make Off With $2.1 Million

Amazon copied products and rigged search results, documents show

5 novembre 2021 /

https://no-flux.beaude.net/wp-content/uploads/2021/11/Amazon.in1_.jpg

“Thousands of pages of internal Amazon documents examined by Reuters – including emails, strategy papers and business plans – show the company ran a systematic campaign of creating knockoffs and manipulating search results to boost its own product lines in India, one of the company’s largest growth markets. The documents reveal how Amazon’s private-brands team in India secretly exploited internal data from Amazon.in to copy products sold by other companies, and then offered them on its platform. The employees also stoked sales of Amazon private-brand products by rigging Amazon’s search results so that the company’s products would appear, as one 2016 strategy report for India put it, “in the first 2 or three … search results” when customers were shopping on Amazon.in.”

Source : Amazon copied products and rigged search results, documents show

Facebook’s facial recognition announcement doesn’t apply to the metaverse

5 novembre 2021 /

Multiple screens showing a human avatar with lines and dots indicating movement vectors.

“Several of Meta’s current projects show that the company has no plans to stop collecting data about peoples’ bodies. Meta is developing hyper-realistic avatars that people will operate as they travel through the metaverse, which requires tracking someone’s facial movements in real time so they can be recreated by their avatar. A new virtual reality headset that Meta plans to release next year will include sensors that track peoples’ eye and facial movements. The company also weighed incorporating facial recognition into its new Ray-Ban smart glasses, which allow the wearer to record their surroundings as they walk around, and Reality Labs, Meta’s hub for studying virtual and augmented reality, is conducting ongoing research into biometrics, according to postings on Facebook’s careers website.”

Source : Facebook’s facial recognition announcement doesn’t apply to the metaverse – Vox

We’re shutting down the Face Recognition system on Facebook | Meta

2 novembre 2021 /

“We’re shutting down the Face Recognition system on Facebook. People who’ve opted in will no longer be automatically recognized in photos and videos and we will delete more than a billion people’s individual facial recognition templates.This change will also impact Automatic Alt Text (AAT), which creates image descriptions for blind and visually-impaired people. After this change, AAT descriptions will no longer include the names of people recognized in photos but will function normally otherwise. We need to weigh the positive use cases for facial recognition against growing societal concerns, especially as regulators have yet to provide clear rules.”

Source : An Update On Our Use of Face Recognition | Meta

The Metaverse Was Lame Even Before Facebook, By Ethan Zuckerman

1 novembre 2021 /

An image of a person wearing VR headset in 1995

“Neal Stephenson’s metaverse has been a lasting creation because it’s fictional. It doesn’t have to solve all the intricate problems of content moderation and extremism and interpersonal interaction to raise questions about what virtual worlds can give us and what our real world lacks. Today’s metaverse creators are missing the point, just like I missed the point back at Ted’s Fish Fry in 1994. The metaverse isn’t about building perfect virtual escape hatches—it’s about holding a mirror to our own broken, shared world. Facebook’s promised metaverse is about distracting us from the world it’s helped break.”

Source : The Metaverse Was Lame Even Before Facebook – The Atlantic

Giant, free index to world’s research papers released online

29 octobre 2021 /

Carl Malamud

“In a project that could unlock the world’s research papers for easier computerized analysis, an American technologist has released online a gigantic index of the words and short phrases contained in more than 100 million journal articles — including many paywalled papers. The catalogue, which was released on 7 October and is free to use, holds tables of more than 355 billion words and sentence fragments listed next to the articles in which they appear. It is an effort to help scientists use software to glean insights from published work even if they have no legal access to the underlying papers, says its creator, Carl Malamud. He released the files under the auspices of Public Resource, a non-profit corporation in Sebastopol, California that he founded. ”

Source : Giant, free index to world’s research papers released online

Facebook files – Le Monde

29 octobre 2021 /

https://no-flux.beaude.net/wp-content/uploads/2021/10/55e1964_874736917-cover-algo.jpg

Facebook : reprendre le contrôle

Ces documents ont levé le voile sur la face sombre de l’entreprise, accusée de privilégier le profit au détriment de sa responsabilité sociétale et démocratique.

Comment l’algorithme de Facebook échappe au contrôle de ses créateurs

Dans des documents internes de l’entreprise, ses ingénieurs avouent leur incompréhension face à un code informatique aux effets imprévus, qui fait du réseau social une machine complexe, difficile à maîtriser.

Sur Facebook, le succès inouï des « fermes à trolls »

« Ce type de pages spécialisées dans la republication de contenus viraux de piètre qualité, pilotées depuis des pays comme l’Albanie ou le Kosovo, touchait plus de 360 millions d’utilisateurs par semaine à la fin de 2019, selon un document interne à l’entreprise ».

Sur Facebook, l’autocensure des utilisateurs les plus modérés en politique

« De multiples études menées par le réseau social montrent que les personnes aux convictions tempérées, de droite comme de gauche, craignent de s’y exprimer sur les sujets politiques, laissant le champ libre aux messages les plus extrêmes ».

Facebook Files : hors des Etats-Unis, les failles de la modération dans des dizaines de langues

« Dans des pays à haut risque, les systèmes de modération du réseau social face aux discours de haine et aux manipulations ne sont pas suffisants, révèlent des documents internes de l’entreprise. »

Sur Facebook, le fiasco de la modération en langue arabe

« Les documents internes de l’entreprise montrent son inertie face aux dysfonctionnements occasionnés par son défaut de maîtrise de la troisième langue du réseau social ».

Des pro-anorexie aux pro-QAnon : Facebook face aux effets pervers de ses recommandations automatiques

« Plusieurs documents internes à l’entreprise pointent les dangers des algorithmes chargés de proposer aux utilisateurs des pages ou groupes qu’ils ne suivent pas ».

La réponse de Facebook au flot de messages antivaccination a été mal ciblée

« D’après des documents internes, le réseau social a vite réagi contre les messages antivaccination. Mais les outils mis en place n’ont pas suffi à endiguer les nombreux messages problématiques ».

Au cœur de Facebook, déprime et frustration dans les équipes chargées de rendre le site plus « sain »

« Les services baptisés « Integrity » détaillent, dans des documents internes, leur manque de moyens et les arbitrages perdus en interne ».

Source : Facebook files – Actualités, vidéos et infos en direct

« Older posts Newer posts »

Archives

Catégories

© 2025 no-Flux

Theme by Anders Noren — Up ↑