IEEE Spectrum Recent Content full text

AI Creates Fake Obama

Charles Q. Choi — Wed, 12 Jul 2017 14:00:00 GMT

Videos of Barack Obama made from existing audio, video of him

Photo: University of Washington.

Artificial intelligence software could generate highly realistic fake videos of former president Barack Obama using existing audio and video clips of him, a new study [PDF] finds.

Such work could one day help generate digital models of a person for virtual reality or augmented reality applications, researchers say.

Computer scientists at the University of Washington previously revealed they could generate digital doppelgängers of anyone by analyzing images of them collected from the Internet, from celebrities such as Tom Hanks and Arnold Schwarzenegger to public figures such as George W. Bush and Barack Obama. Such work suggested it could one day be relatively easy to create such models of anybody, when there are untold numbers of digital photos of everyone on the Internet.

The researchers chose Obama for their latest work because there were hours of high-definition video of him available online in the public domain. The research team had a neural net analyze millions of frames of video to determine how elements of Obama's face moved as he talked, such as his lips and teeth and wrinkles around his mouth and chin.

In an artificial neural network, components known as artificial neurons are fed data, and work together to solve a problem such as identifying faces or recognizing speech. The neural net can then alter the pattern of connections among those neurons to change the way they interact, and the network tries solving the problem again. Over time, the neural net learns which patterns are best at computing solutions, an AI strategy that mimics the human brain.

In the new study, the neural net learned what mouth shapes were linked to various sounds. The researchers took audio clips and dubbed them over the original sound files of a video. They next took mouth shapes that matched the new audio clips and grafted and blended them onto the video. Essentially, the researchers synthesized videos where Obama lip-synched words he said up to decades beforehand.

The researchers note that similar previous research involved filming people saying sentences over and over again to map what mouth shapes were linked to various sounds, which is expensive, tedious and time-consuming. In contrast, this new work can learn from millions of hours of video that already exist on the Internet or elsewhere.

One potential application for this new technology is improving videoconferencing, says study co-author Ira Kemelmacher-Shlizerman at the University of Washington. Although teleconferencing video feeds may stutter, freeze or suffer from low-resolution, the audio feeds often work, so in the future, videoconferencing may simply transmit audio from people and use this software to reconstruct what they might have looked like while they talked. This work could also help people talk with digital copies of a person in virtual reality or augmented reality applications, Kemelmacher-Shlizerman says.

The researchers note their videos are currently not always perfect. For example, when Obama tilted his face away from the camera in a target video, imperfect 3-D modeling of his face could cause parts of his mouth to get superimposed outside the face and onto the background.

In addition, the research team notes their work did not model emotions, and so Obama's facial expressions in the output videos could appear too serious for casual speeches or too happy for serious speeches. However, they suggest that it would be interesting to see if their neural network could learn to predict emotional states from audio to produce corresponding visuals.

The researchers were careful to not generate videos where they put words in Obama's mouth that he did not at some other time utter himself. However, such fake videos are “likely possible soon,” says study lead author Supasorn Suwajanakorn, a computer scientist at the University of Washington.

However, this new research also suggests ways to detect fake videos in the future. For instance, the video manipulation the researchers practiced can blur mouths and teeth. “This may be not noticeable by human eyes, but a program that compares the blurriness of the mouth region to the rest of the video can easily be developed and will work quite reliably,” Suwajanakorn says.

The researchers speculated that the link between mouth shapes and utterances may be to some extent universal for people. This suggests that a neural network trained on Obama and other public figures could be adapted to work for many different people.

The research was funded by Samsung, Google, Facebook Intel and the University of Washington. The scientists will detail their findings [PDF] on Aug. 2 at the SIGGRAPH conference in Los Angeles.

Iran's Newest Robot Is an Adorable Dancing Humanoid

Erico Guizzo — Wed, 12 Jul 2017 13:12:00 GMT

University of Tehran roboticists have built a dancing, karate-chopping little humanoid called Surena Mini

Photo: University of Tehran

Iranian researchers have recently unveiled a new robot called Surena Mini.

Over the last several years, a team of roboticists at the University of Tehran has been working on increasingly large and complex life-size humanoids. For their latest project, however, the Iranian researchers decided to build something smaller—and cuter.

Surena Mini is a knee-high robot with a sleek 3D-printed body, articulated limbs, and a round head with two camera-eyes. Twenty small servomotors power its arms, legs, and neck, allowing the little robot to walk, gesture, and dance:

“The main purpose of this robot is to provide researchers and students with a reliable robotic platform for educational and research applications,” Aghil Yousefi-Koma, a professor of mechanical engineering at the University of Tehran, told IEEE Spectrum.

He added that his group also has plans to offer the robot “for helping autistic and deaf children.”

A team of 15 researchers at University of Tehran’s Center for Advanced Systems and Technologies worked for over a year to design and build Surena Mini, which is 50 centimeters tall and weighs 3.4 kilograms.

Packed inside the robot are a compact computer with an Intel Core CPU, cameras and infrared sensors, speakers, and an IMU, or inertial measurement unit. Its hands aren’t designed for grasping objects, but Surena Mini can push on small things—or karate-chop them:

A little over a year ago, the same group unveiled Surena III, an advanced adult-size humanoid designed for researching bipedal locomotion, human-robot interaction, and other challenges in robotics.

Surena III, equipped with cameras, 3D sensor, and a computer running ROS, or Robot Operating System , was able to pick up bottles, imitate a person’s gestures, and stand on one foot.

Photos: University of Tehran

Iranian researchers unveiled Surena III in 2015. The robot, almost 2 meters tall and weighing 98 kilograms, can kick a ball, go up a ramp, and walk down a set of steps.

The Iranian roboticists plan to continue working on Surena III, but they also want to explore the possibility of creating marketable products based on their research, Professor Yousefi-Koma explained, and one of the ideas they had was building a “kid-size version of Surena.”

Surena Mini’s overall size and design appear similar to that of other small humanoids like Nao, developed by French robotics company Aldebaran (now SoftBank Robotics), and Robotis OP2, created by U.S. and South Korean roboticists.

But the Iranian robot has yet to show that it has some of the same skills already demonstrated by those other humanoids. Nao and Robotis OP2 have been used in research labs, schools, and hospitals for nearly a decade. Both are also used in the RoboCup robot soccer competition.

Photo: University of Tehran

Researchers from the University of Tehran’s Center for Advanced Systems and Technologies, led by Professor Aghil Yousefi-Koma (standing between the robots with red and green feet/hands), worked for over a year to design and build Surena Mini.

Despite their size, these little robots are pricey. Nao and Robotis OP2 each sell for nearly US $10,000. Professor Yousefi-Koma said Surena Mini will be available for 260,000,000 Iranian rials, or $8,000, but he hopes the cost to come down if the robot can be produced in large batches.

One of the biggest challenges of the project, he explained, has been implementing features like face detection and voice recognition, which would let the robot perform with a greater level of autonomy. His team has developed such capabilities for their large robots, but replicating them using Surena Mini’s limited hardware is a trickier task.

To program the robot, advanced users can modify the source code to create different behaviors. But the researchers wanted to make Surena Mini accessible to less experience users as well. So they created a programming environment with a graphical interface “designed to be attractive and user-friendly,” Professor Yousefi-Koma said.

“It gives users full access to all available features of the robot,” he added, “so even beginners can make the robot walk and move its arms and head.”

The Audi A8: the World's First Production Car to Achieve Level 3 Autonomy

Philip E. Ross — Tue, 11 Jul 2017 17:00:00 GMT

It's also the first to sport lidar

The 2018 Audi A8, just unveiled in Barcelona, counts as the world’s first production car to offer Level 3 autonomy.

Level 3 means the driver needn’t supervise things at all, so long as the car stays within guidelines. Here that involves driving no faster than 60 kilometers per hour (37 mph), which is why Audi calls the feature AI Traffic Jam Pilot.

Go ahead, Audi’s saying, read your newspaper or just zone out while traffic creeps along. Take a look at the company’s promotional video. Beginning around the 3:55 mark you’ll see an indulgent father who ends up horsing around with his kid in the back seat. When the car up ahead stops, the A8’s AI hits the brakes in time to avoid rear-ending it.

Other cars now on the market may physically allow the driver to zone out, too, but not without squawking a lot. “Put your hand on the wheel within five seconds, Buster, or I’m pulling over to the side of the road,” they’ll say, in so many words.

Mercedes-Benz sounds an alarm if you take your hand off the wheel. Tesla makes you hit the turn signal to show you want to change lanes. The new Cadillac CT6 goes so far as to monitor your eye movements with infrared cameras to make sure you’re paying attention to the road. That’s an emphatic statement of the car’s status: Level 2.

To be sure, the A8 also monitors the driver, even while the traffic jam persists, and continues to do so as the speed edges up over the limit. If the driver falls asleep, it’ll wake him up; if it can’t get his attention, it will stop the car.

There’s no one feature that seems to be behind the company’s decision to go up to Level 3, but there are certainly a lot of new technologies. There are computers from Nvidia and other firms, an image processor from Mobileye, and a really huge array of sensors: 12 ultrasound sensors, five cameras, five radars, one infrared camera for night vision. Most notable of all, there’s lidar—the first ever offered on a production car. The unit, a forward-looking one, comes from Valeo.

Besides allowing for Level 3 autonomy, the panoply of devices also makes for a smoother ride. For instance, when the sensors see a pothole coming up they prime the active suspension so that it can handle the challenge more easily. Oh, and the car is also a mild hybrid, with a plug-in version in the works.

Electric-drive is suddenly catching on—just last week Volvo announced that all its 2019 cars would have at least some degree of electric drive in them. Nowhere has the change in attitude been more extraordinary than in the family of VW, Audi’s parent company, which had based its emissions policy on what it called clean diesel. VW had to give it up last year, when it was shown to have cheated on diesel emissions tests.

If you want to buy the new A8, you’ll have to check whether your jurisdiction will accept it as a Level 3 car. Audi said in a statement that it will follow “a step-by-step approach” to introducing the traffic jam pilot. It plans to sell the base model in Europe this fall for €90,600, or about US $103,000, and to enter the United States market shortly afterwards. A model having a longer wheelbase will cost a few percent more.

What’s next? A Level 4 car, naturally. Back in January, Audi and Nvidia said they’d have one on the roads by 2020.

NATO Unveils JANUS, First Standardized Acoustic Protocol for Undersea Systems

Amy Nordrum — Tue, 11 Jul 2017 16:30:00 GMT

The new acoustic communications protocol is a step towards an Internet of Underwater Things

Photo: NATO

Aquatic robots are busier than ever. They have seabeds to mine, cable pathways to plough, and marine data to gather. But they and their aquatic brethren—including submarines and scuba divers—still struggle to communicate.

For decades, global standards defining Wi-Fi and cellular networks have allowed people to exchange data over the air. But those technologies are worthless below the waves, and no such standards have existed for underwater communications.

Aquatic systems have instead used a mishmash of acoustic and optical signals to send and receive messages. However, manufacturers sell acoustic modems that operate at many different frequencies, which means those systems often can’t speak to each other.

“We live in a time of wild west communications underwater,” says João Alves , a principal scientist for NATO.

Now, Alves and other NATO researchers have established the first international standard for underwater communications. Named JANUS , after the Roman god of gateways , it creates a common protocol for an acoustic signal with which underwater systems can connect.

Acoustics has long been a popular medium for underwater communications. Generally, optical signals can deliver high data rates underwater at distances up to 100 meters, while sound waves cover much greater distances at lower data rates.

The main role of JANUS is to bring today’s acoustic systems into sync with one another. It does this in part by defining a common frequency— 11.5 kilohertz —over which all systems can announce their presence. Once two systems make contact through JANUS, they may decide to switch to a different frequency or protocol that could deliver higher data rates or travel further.

In this way, Alves compares JANUS to the English language—two visitors to a foreign country may speak English to one another before realizing they are both native Spanish speakers, and switch to their native tongue.

Photo: Hans Camenzind

IEEE Spectrum Recent Content full text

AI Creates Fake Obama

Iran's Newest Robot Is an Adorable Dancing Humanoid

The Audi A8: the World's First Production Car to Achieve Level 3 Autonomy

NATO Unveils JANUS, First Standardized Acoustic Protocol for Undersea Systems

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Intersil ICL8038 Waveform Generator

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: STMicroelectronics STA2056 GPS Receiver

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Texas Instruments Digital Micromirror Device

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Intel 8088 Microprocessor

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Western Digital WD1402A UART

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Toshiba NAND Flash Memory

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Transmeta Corp. Crusoe Processor

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Texas Instruments TMS32010 Digital Signal Processor

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Kodak KAF-1300 Image Sensor

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Computer Cowboys Sh-Boom Processor

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Micronas Semiconductor MAS3507 MP3 Decoder

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Amati Communications Overture ADSL Chip Set

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Sun Microsystems SPARC Processor

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Microchip Technology PIC 16C84 Microcontroller

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Texas Instruments TMC0281 Speech Synthesizer

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Fairchild Semiconductor μA741 Op-Amp

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame

Chip Hall of Fame: Acorn Computers ARM1 Processor

< Back to the Chip Hall of Fame

< Back to the Chip Hall of Fame