Goto

Collaborating Authors

Results


Facebook's Dynabench 'Radically Rethinks AI Benchmarking'

#artificialintelligence

In the ever-expanding world of computer hardware and software, benchmarks provide a robust method for comparing quality and performance across different system architectures. From MNIST to ImageNet to GLUE, benchmarks have also come to play a hugely important role in driving and measuring progress in AI research. When introducing any new benchmark, it's generally best not to make it so easy that it will quickly become outdated, or so hard that everyone will simply fail. When new models bury benchmarks, which is happening faster and faster in AI these days, researchers must engage in the time-consuming work of making new ones. Facebook believes that the increasing benchmark saturation in recent years -- especially in natural language processing (NLP) -- means it's time to "radically rethink the way AI researchers do benchmarking and to break free of the limitations of static benchmarks." Their solution is a new research platform for dynamic data collection and benchmarking called Dynabench, which they propose will offer a more accurate and sustainable way for evaluating progress in AI.


Tracking Recent Topics & Trends in 3D Photo Generation

#artificialintelligence

Today any smartphone can generate 3D Photos, but the popular AI-powered effect is actually fairly new. It was back in 2018 that Facebook first introduced a machine learning-based 3D photo feature that allowed users to generate an immersive 3D image from normal 2D pictures. Leveraging the dual-lens "portrait mode" capabilities that had recently become available in smartphones, the feature quickly gained traction and began evolving. This June, a research team from Virginia Tech, National Tsing Hua University and Facebook designed an algorithm that generates even more immersive 3D photos from a single RGB-D (colour and depth) image. And in August, Facebook democratized the technique with a novel system able to generate 3D photos even on low-end mobile phones or without an Internet connection. Facebook isn't the only tech giant using AI to generate 3D photos -- in recent months, Google has introduced its own AI techniques for generating 3D photos from 2D images.


LinkedIn's 50 Best Startups To Work For In 2020

#artificialintelligence

These and other insights are from LinkedIn's Top Startups 2020: The 50 U.S. companies on the rise published today. This is the 4th annual LinkedIn list of the hottest startups to work for. The list is determined by the billions of actions taken by LinkedIn's 706 million members. The annual list is a reflection of how business and work is evolving through the pandemic, what industries are emerging and growing and where people want to work now, reflecting the current state of the economy and the world. Even in the face of Covid-19, the startups on this year's list are all still innovating and experiencing growth and the majority of the companies on the list are currently hiring, with 3,000 jobs now open on LinkedIn. To be eligible for the list, a company must be independent and privately held, have at least 50 employees, be seven years old or younger, be headquartered in the country on the list which they appear and have a minimum of 15% employee growth over the time period. The top 50 U.S. startups include the following: Full-time headcount: 4,000 Headquarters: New York City Year founded: 2016 What you should know: While the U.S. economy quickly sank into a recession at the start of the pandemic, one of its engines has been roaring: housing.


Facebook AI Enables Automatic Closed Captions for Facebook Live

#artificialintelligence

Conventional ASR systems are generally made up of three components: an acoustic model that predicts phonemes from short segments of audio, a pronunciation lexicon which describes how the phonemes are combined to form the words in a given language, and a language model that captures the relationships among those words. Facebook engineers have deployed their model variations with a number of infrastructure optimizations to handle the additional livestream traffic while also reducing the compute required despite the increased load. Although the system was trained on many different types of speech, it's still far from perfect, particularly when it comes to accents. As it's difficult to collect sufficient training data for every accent type, Facebook researchers are now exploring ways to improve their models by having them also learn from the vast amounts of unlabelled audio that is available online.


Tinder launches apocalyptic Swipe Night experience in the UK and around the world

Mashable

Trying to find love as the world ends? That premise is central to Tinder's interactive Swipe Night event, which launches in the UK and around the world on Sept. 12 at 10am. If you're unfamiliar with Swipe Night, then here's a lil catch up: Swipe Night is a first-person choose-your-own-adventure style event where Tinder users can swipe at key moments to determine the direction of the story within the app. Swiping doesn't just affect how the story ends -- it also has a bearing on who users match with and what they end up chatting about. As for the storyline, well, it couldn't be more pertinent to the times we're living in.


Singles don't want to date non-voters, according to new OkCupid survey

Mashable

If I'm perusing a dating app and someone mentions being apolitical, or not caring about politics, I grimace. In this (ravaged) economy (and global pandemic and time of social unrest)? I'm not alone in this, according to data found by OkCupid. Over 500,000 users said they couldn't date someone who didn't vote, according to new data provided by the dating app. Those who say they're registered voters are 63 percent more likely to get a match -- and 85 percent more likely to receive a message.


To get skin tones right, Pinterest's AI went way beyond facial recognition

#artificialintelligence

"Before we launched skin tone ranges in 2018, nearly sixty percent of the top 100 search terms for skin-related searches involved a tone, such as dark, pale, and olive tones, showing people wanted a way to customize their searching," wrote Nadia Fawaz, Pinterest's technical lead for fairness in AI, in an email to Fast Company. Now, the company has unveiled a new version of its technology designed to more accurately detect skin tones across a broader class of images, quadrupling the number of beauty and fashion pinned items where its algorithms can spot a skin tone. Pinterest's skin-tone AI is used in a range of features letting people use the platform to find custom looks for themselves, including a recently unveiled augmented-reality Try On tool that lets people use their smartphone's camera to see how various lip colors look on them. That's good for Pinterest users who are using the platform to shop--and, of course, for Pinterest advertisers participating in the Try On program. The company claims millions of users come each month seeking beauty ideas.


3 Daunting Ways AI Will Transform The World Of Work

#artificialintelligence

Each industrial revolution has brought with it new ways of working – think of the impact computers and digital technology (the third industrial revolution) have had on how we work. But this fourth industrial revolution – what I call the "intelligence revolution," because it is being driven by AI and data – feels unprecedented in terms of the sheer pace of change. The crucial difference between this and the previous industrial revolutions is we're no longer talking about generational change; we're talking about enormous transformations that are going to take place within the next five, 10 or 20 years. Here are the three biggest ways I see AI fundamentally changing the work that humans do, within a very short space of time. Increasing automation is an obvious place to start since a common narrative surrounding AI is "robots are going to take all our jobs."


Match Group Looks to Capitalize on Video Dating During the Pandemic

WSJ.com: WSJD - Technology

Dallas-based Match operates several dating apps, including Tinder, Hinge and OkCupid, as well as its namesake brand. The company in July completed its separation from IAC/InterActiveCorp., which previously owned a roughly 80% stake. Match released video-chatting features for its apps in the spring as users started avoiding traditional dating spots such as bars and restaurants. The company is now in the beginning stages of developing features such as games and icebreakers to make those one-on-one video calls more engaging--part of a broader strategy to find new ways to generate revenue from its millions of users, according to Chief Financial Officer Gary Swidler. "We've got a lot of users, and I think there's more we can do with them," said Mr. Swidler, who is also Match's chief operating officer.


The Newest Listed Stock in the $100B USD Digital Ad Space

#artificialintelligence

Nowhere is more apparent than the massive growth within the online marketing advertising sector. The effects of lockdowns and quarantines forcing companies to operate remotely and keeping people at home shone a light on increased software use, digital reliance, and online advertising. Before the pandemic, the largest companies in the world were already tech giants. All of the big five have fought for and earned revenue from software services, but the growing advertising slice of the pie had become dominated by the duopoly of Facebook (NASDAQ:FB) and Alphabet (NASDAQ:GOOG). Now, that media advertising market has grown to US $100 billion largely on advertising revenue and digital ad spending that are both forecasted to keep increasing.