Skip to main content

3D News

Dear 3D Vision Fans, We have enjoyed sharing 3D Vision content here over the past eight years, and seeing the fantastic images and videos you’ve uploaded. Thanks for making this such a fun destination for 3D enthusiasts. Usage of this site has been declining, and therefore, we are going to transition the 3D photo function of this site over to Phereo.com. The last day that you will be able to access 3DVisionLive.com photos is Aug. 1, 2018. In addition, the 3D video...

Recent Blog Entries

The dynamics of racing — a human driver guiding a powerful machine to reach the finish line as fast as possible — haven’t changed much since the Goodwood Festival of Speed was founded a quarter century ago. Until now.

The Roborace Robocar, a futuristic vehicle designed for the new era of driverless transportation, became the first car to complete the Goodwood hillclimb without a human at the wheel. The innovative race car, powered by the NVIDIA DRIVE autonomous vehicle computing platform, wowed 55,000 spectators by navigating with ease around trees, hay bales and tight curves.

“We are ecstatic that the team have been able to achieve this landmark run and we hope that it draws attention to the amazing advances that are being made in the automotive industry,” said Rod Chong, deputy CEO of Roborace.

The Goodwood Festival is an annual weekend-long celebration of racing, with vehicles ranging from classic cars to Formula 1 race cars running their paces in front of a huge crowd. The feature event, known as the hillclimb, is a 1.16-mile course with a twisting, steady uphill grade, climbing nearly vertical 500 feet. This race is no Sunday drive — racing enthusiasts describe hillclimbs as the combination of auto racing and mountain climbing. Competitors must carefully navigate bumpy courses and maintain grip on slippery terrain, all while going as fast as possible.

The Roborace Robocar, designed by automotive futurist Daniel Simon — known for his work in Hollywood films such as Oblivion and Tron: Legacy — is a crouching sleek racer made to be driven by artificial intelligence. The car has no seat, steering wheel or pedals, but packs a punch — four 135kW electric motors power each wheel for a combined 500-plus horsepower.

With the NVIDIA DRIVE platform under the hood, the Robocar can pilot itself autonomously by processing data from camera, radar and lidar sensors in real time. Using that data, the vehicle is able to determine its location on the hill and detect areas where it can drive on a course surrounded by trees and other obstructions.

Taking a human out of the driver’s seat is a significant feat, not to mention doing so at one of the summer’s largest racing event. However, the car didn’t disappoint, sailing through the uphill climb with a maximum speed of 75 mph.

“Robocar is an ambassador for the future technologies we will see on our roads and we hope that inspirational events like this will change public perceptions of autonomous vehicles,” Chong said.

The historical achievement is a significant milestone toward Roborace’s ultimate goal: establishing a new racing platform without human drivers. Competitors will race software installed in the Robocar, testing the durability and improvisation ability of AI.

Roborace will speak about its progress in autonomous racing at GTC Europe in October.

The post Robocar Climbs to New Heights at Goodwood Festival of Speed appeared first on The Official NVIDIA Blog.

It’s possible to get grad-school credit for writing memes. At least if you use deep learning to do it.

Just ask Lawrence Peirson.

The 23-year-old is pursuing a theoretical astrophysics Ph.D. at Stanford, but decided to enroll in a couple AI courses this year. He and classmate E. Meltem Tolunay came up with a neural network that captions memes for a class project, now published in a whitepaper aptly titled “Dank Learning.” (“Dank,” for the uninitiated, is a synonym for “cool.”)

There are lots of examples of training deep learning models to produce literal captions for an image — for example, accurately captioning an image as “man riding a surfboard” or “child with ice cream cone.” With memes, Peirson’s challenge was to see if a neural network could go beyond literal interpretation and create humorous captions.

Though he was initially skeptical that the memes would be funny, Peirson found that the deep learning model produced “some quite interesting and original humor.”

Attaining Deep Meme

The deep learning network captioned this meme, a variation on the popular advice animals template.To collect training data for the deep learning model, Peirson scraped around 400,000 user-generated memes from the website memegenerator.net. The site provides meme templates and allows users to come up with their own captions.

The dataset included around 3,000 base images, each with many different captions. Since the input data was crowdsourced, there was a wide range in quality of meme captions the deep learning model processed.

“With 400k memes, most aren’t going to be that funny, but at least they teach the system what a meme is, what joke is relevant,” he said.

Internet memes have circulated around the web for years, with a strong foothold in websites like Reddit, Facebook, 9GAG and Quick Meme. The most popular can get more than 2 million unique captions created.

Memes often reference pop culture, current events or esoteric bits of a particular internet subculture. (Peirson runs a meme page called “The specific heat capacity of europium at standard temperature and pressure.”)

As a result, they imbibe both the good and bad of digital culture — the paper notes a bias in the training data towards expletive, racist and sexist memes. Peirson sees the need to filter these out in future work, but points out that “it’s a big problem in natural language processing in general,” not one specific to memes.

The deep learning model was programmed in CUDA and used an NVIDIA TITAN Xp GPU. Peirson and Tolunay tried using both unlabeled data and data labeled with the meme title (for example, success kid or trollface), but saw no significant difference in meme quality.

“They’re very funny in a ‘it sort of makes sense, but not really’ way,” Peirson said. “Memes lend themselves to that kind of humor.”

The deep learning network captioned this meme, a variation on the popular advice animals template. One Does Not Simply Declare a Meme Dank

To evaluate the deep learning model’s success, the collaborators calculated a perplexity score, which checks whether the neural network can identify clear patterns in the data. They calculated this metric for a few hundred memes with preset formats, such as the Boromir meme, which always begins with the phrase “one does not simply.”

But the true test of a meme is whether it’s funny.

In a qualitative survey, Peirson and his co-author presented people with a human-generated and a deep learning-generated meme, side by side. They asked two questions: whether the subject thought the meme was created by a human or computer, and how the subject would rate the meme’s humor.

The data shows the deep learning memes “were fairly indistinguishable from the real memes,” Peirson says.

They also investigated how the neural network would caption images that were not among the templates in the training dataset. In these cases, the algorithm infers patterns from the unknown image based on what it has seen in the training data. Peirson even showed the deep learning system a photo of his own face to test this, with entertaining results.

When Peirson ran the deep learning model on a photo of his own face, this is one of the captions it came up with.

Memes often go viral, and it seems meme-themed whitepapers are no exception. Peirson says he was “extremely surprised” by the media coverage and wide interest in the project. A complementary mobile app, also titled Dank Learning, will soon be available on the App Store.

This project, he says, has given him a new perspective on how powerful memes can be. Millions of users worldwide consume memes daily on social media sites.

Peirson sees the potential for a strong AI to produce memes “at a whim” on current events to influence public opinion — or for advertisers to use memes for brand awareness: “Having this go viral is an incredible way to market.”

The post Living the Meme: AI as Funny as Humans for Generating Image Captions appeared first on The Official NVIDIA Blog.