Facebook’s sticker search has a queasy political reaction
I don’t usually search out particular stickers to express myself on Facebook Messenger (I’m a 💯 or 🔥 emoji kind of guy) but recently people who do, have noticed some strange results. VentureBeat points out a tweet by Serena Ehrlich showing the vomit sticker in response to searches for liberals or feminism. The company responded saying something in its search algorithm caused the response and that it should not continue to appear after today.
Whatever caused it, the reaction wasn’t entirely partisan, appearing in response to searches for Trump, Obama (but not Barack Obama), Biden (but not Joe Biden) and politics in general, although several conservative-linked terms that I tried pulled up nothing at all. Based on those results, it’s hard to say if this is another algorithmic misstep, the result of synthesizing this election season’s conversations or the tagging work of pranksters. No matter what the cause, it will be gone soon, and peace will once again reign over the land of messaging.
Source: VentureBeat
Facebook promises to stabilize shaky 360-degree videos
Shaky video is already a problem with conventional cameras, but it’s much worse with 360-degree and virtual reality cams. Your bumpy mountain bike ride won’t be so exciting to watch if it makes viewers queasy. However, Facebook might just save the day: it’s testing an algorithm designed to stabilize 360-degree shots. The approach blends 2D motion models with 3D reconstruction to reduce the unwanted effects in immersive footage, such as the bobbing camera movements, lens deformations and stitching between cameras.
It’s reportedly more accurate thanks to the 3D elements, but faster and more “robust” through its partial use of 2D techniques. Moreover, it’s efficient. You can stabilize a video faster than it takes to play it, and you can reduce the bitrate for a video by up to 20 percent. That, in turn, means smaller file sizes and more reliable video streaming.
Facebook has some work to do before the stabilization code is ready, but it plans to make the technology available to everyone on both Facebook and Oculus devices. It won’t be ubiquitous, but you’ll definitely notice its presence. If you want software to play with, though, Facebook isn’t leaving you empty-handed. It’s open-sourcing Zstandard, a data compression algorithm designed to scale with new computer hardware, as well as its space-savvy MyRocks MySQL database.
Via: TechCrunch
Source: Facebook Code (1), (2), (3)
Instagram gets pinch-to-zoom on iOS, Android update in tow
The feature you didn’t know you wanted for Instagram has finally arrived: pinch-to-zoom. Surprised it wasn’t already there? Don’t be. Originally, Instagram didn’t support large enough images to warrant enlarging — but last year, that changed last year when the company bumped its default image size up to 1080 x 1080. Now, we’re seeing the fruits of that upgrade: starting today, iOS users can pinch and zoom any image or video in their feed. On Android? Sit tight. Instagram says the update will roll out to additional platforms in a few weeks.
A video posted by Instagram (@instagram) on Aug 31, 2016 at 8:00am PDT
Source: Instagram
Chris Brown turns to Instagram amid police standoff
Singer Chris Brown’s Instagram account got a lot more interesting today. With police outside of his Los Angeles home, who were following up on a report that Brown pulled a gun on an unnamed woman, the singer started posting Instagram videos criticizing the LAPD. Meanwhile, local news stations have been broadcasting footage of the standoff on Facebook Live for much of the day.
It’s not unusual for local news stations to turn situations like this into a media circus, but this is a rare instance where we actually got to hear from the person who’s actually being pursued by the police. In the Instagram posts, Brown evoked the Black Lives Matter movement and painted himself as a victim of police harassment. “I don’t care y’all going to stay playing with me like I’m the villain out here, like I’m going crazy,” he said.
“Right now we’re getting cooperation from everybody who’s involved in this,” an LAPD representative said during a news briefing. The police are currently evacuating the home, after issuing a search warrant, and are interviewing everyone who was inside.
Via: Fox News, TMZ
Source: Chris Brown (Instagram)
Instagram Stories fights Snapchat by recommending users to follow
Instagram Stories’ feature updates are no longer quite in lockstep with Snapchat’s. It’s gradually rolling out an update to the Explore section that displays a bar of Stories from people Instagram thinks you want to follow based on both your existing contacts and your favorite topics. The addition could give you extra Stories clips to watch even when your usual Instagram friends aren’t up to snuff. The concept isn’t entirely new (Snapchat’s featured Stories are slightly similar), but it’s uniquely tailored to you — you’re not just getting the same editor’s picks as everyone else.
The seemingly simple tweak could be important to keeping Instagram Stories in the limelight and reducing Snapchat’s first-mover advantage. Instagram tells our TechCrunch friends that 100 million people check out the Explore tab every day. When the social service has over 500 million people, that’s a large chunk of the audience seeing videos they’d otherwise have missed. Only a fraction of users are likely to follow others just for the sake of their Stories, of course, but that could be enough to keep you interested in the feature after the novelty wears off.
Source: TechCrunch
Zuckerberg hopes to show off his home control AI next month
Remember that artificial intelligence Mark Zuckerberg said he was going to build to control his house? It sounds like he actually made it — and he’s almost ready to show it to the public. “I got it to the point where I can control the lights, I can control the gates, I can control the temperature,” he said at a Facebook Q&A in Rome this week. “It’s getting there.”
Zuckerberg says he hopes to have a demo of the project available sometime next month, and credits a lot of the AI’s success to his engineering team at work. “It’s awesome because I get to interact with all these Facebook engineers who are doing this awesome AI work in speech recognition, in face recognition,” he says. “I programmed it so now, when I walk up to my door, my gate, I don’t have to put in a code or something like that to get in — put in a key — it just sees my face and lets me in.” Even so, the AI isn’t quite ready for primetime: Zuckerberg says the house only responds to his own voice, much to his wife’s disappointment. “I’ll give her access once I’m done.” Sure you will, Zuck. Sure you will.
Via: Verge
Source: Facebook
Facebook is working on user-activated Safety Checks
Facebook’s Safety Check has proven an invaluable tool for people to contact their friends and families in the immediate aftermath of large scale disasters. At a public Q&A session in Luiss University in Rome on Monday, CEO Mark Zuckerberg told the crowd that his company is also working on a means of letting any user activate the emergency system on their own.
“If we’re building a community product, this is one of the moments of truth for us,” Zuckerberg explained. “When Safety Check got started a couple of years ago, it was only for natural disasters. Unfortunately since then we’ve had to expand it to terrorist attacks too, because that’s just been too common over the last few years. The next thing we need to do is make it so that communities can trigger it themselves when there is some disaster.”
There’s no word yet as to when the feature will actually go live, nor did Zuckerberg expound on how users would be able to activate it or what events would qualify for a Safety Check.
Source: Venture Beat
Facebook highlights false news story in its trending topics
The usefulness of Facebook’s “trending topics” feature has always been a bit questionable, but it’s been under fire this year as the company battled accusations its human editors were filtering out “conservative” news stories. Today, Facebook is getting a bit more egg on its face: its trending topics area promoted a story about Fox News anchor Megyn Kelly being fired for supporting Hillary Clinton. As of this writing, that has not happened — but the false story was still trending for hours on Facebook before finally getting pulled.
Megyn Kelly is trending on Facebook for an article that has no basis in reality. pic.twitter.com/31f4ERnzHI
— Kyle Blaine (@kyletblaine) August 29, 2016
This comes just a few days after Facebook announced that humans would no longer write descriptions for trending topics, instead letting an algorithm pull relevant excerpts of the articles to display beneath the topic. Despite that, humans are still responsible for deciding what topics are worthy of hitting Facebook’s trending area, but an individual user’s experience will vary based on a number of factors.
If humans are still responsible for overseeing what topics show up in the trending area, though, it’s not great that an entirely false story was promoted on the site for hours before being pulled. On the other hand, even if the story was false, it appears that lots of Facebook users were discussing it — so by that metric, Megyn Kelly was most definitely a trending topic today. But a fake news story certainly does not meet the quality bar that Facebook aspires to, so it’s not surprising the topic has been pulled. Now, even searching “Megyn Kelly” on Facebook doesn’t pull up any stories related to this morning’s fake news.
Update, 12:05PM ET: A Facebook spokesperson said that the trending team accepted the topic over the weekend, because there was a sufficient number of relevant articles and posts, but upon review today they decided to take it down. That decision was made based on the inaccuracy of the story. Facebook didn’t expand on how the story got accepted in the first place, but the company did say its working on improving its detection of hoax and satirical stories.
.@facebook still has a totally fake story about Megyn Kelly in its trending list. Niiceee pic.twitter.com/n2egP2d92h
— Justin Green (@JGreenDC) August 29, 2016
Via: The Verge
UK: Facebook, Google, Twitter ‘consciously failing’ on terrorism
The UK parliament has slammed Facebook, Twitter and YouTube for “consciously failing” to remove terrorism recruitment content. According to a report from the Home Affairs Committee, the social networks are “the vehicle of choice in spreading propaganda and the recruiting platforms for terrorism.” In statements to the WSJ, the companies denied that they are lax with extremist postings. “We deal swiftly and robustly with reports of terrorism-related content,” a Facebook spokesperson said.
The committee based its report on statements from intelligence groups, the Muslim community, counter-terrorism experts and security specialists. Other experts told the WSJ that the document is misleading, saying terrorists are more likely to recruit via heavily encrypted messaging services like WhatsApp and Telegram — apps that are also in the US and UK governments’ crosshairs.
These companies have teams of only a few hundred employees to monitor networks of billions of accounts and Twitter does not even proactively report extremist content to law enforcement agencies.
The role of online networks in abetting terrorism has been a hot topic of late — Twitter said it recently suspended 360,000 terrorism-related accounts since the beginning of the year. The lawmakers think that, given their billions in revenues, the firms aren’t doing enough, though. “These companies have teams of only a few hundred employees to monitor networks of billions of accounts and Twitter does not even proactively report extremist content to law enforcement agencies,” the report states.
The Home Affairs committee wants social websites to take a “zero tolerance approach to online extremism,” and recommended laws that would force social networks to quickly remove terrorist propaganda and inform law enforcement. (While the group’s rulings are non-binding, they heavily influence UK’s parliament.) The European Union recently secured commitments from Facebook, Twitter, Microsoft and Google to put hate speech policies in place and remove and report content within 24 hours.
Source: Parliament.uk
Facebook will show bigger vertical videos in your News Feed
Facebook’s News Feed for mobile will become much friendlier to vertical videos in the near future, according to Marketing Land. It won’t exactly be optimized for the orientation the way Snapchat is, but it will apparently stop cropping and showing them as tiny squares. The publication says when the update rolls out for Android and iOS, you’ll start seeing vertical videos with a 2:3 aspect ratio (as opposed to 1:1) on your News Feed without having to expand them. “We know that people enjoy more immersive experiences on Facebook, so we’re starting to display a larger portion of each vertical video in News Feed on mobile,” a spokesperson told Marketing Land.
Thanks to the popularity of apps like Snapchat, Periscope and Meerkat, more and more people have learned to embrace the format. Daily Mail North America’s CEO Jon Steinberg once said that the publication’s vertical video ads have nine times more completed views than ones shot in horizontal view. By showing a larger part of vertical videos than before, people too lazy to view them in full screen — and, let’s face it, it’s a hassle navigating away from the News Feed sometimes — are more likely to watch them till the end. Facebook didn’t reveal when the feature will go live, but Mashable said the update’s going out in the coming weeks.
Via: Mashable
Source: Marketing Land
AIVAnet 