[I]nside of a Google server or a Facebook server is a little voodoo doll, avatar-like version of you […] All I have to do is simulate what conversation the voodoo doll is having, and I know the conversation you just had without having to listen to the microphone.

{ Quartz | Continue reading }

…a phenomenon privacy advocates have long referred to as the “if you build it, they will come” principle — anytime a technology company creates a system that could be used in surveillance, law enforcement inevitably comes knocking. Sensorvault, according to Google employees, includes detailed location records involving at least hundreds of millions of devices worldwide and dating back nearly a decade.

The new orders, sometimes called “geofence” warrants, specify an area and a time period, and Google gathers information from Sensorvault about the devices that were there. It labels them with anonymous ID numbers, and detectives look at locations and movement patterns to see if any appear relevant to the crime. Once they narrow the field to a few devices they think belong to suspects or witnesses, Google reveals the users’ names and other information. […]

Google uses the data to power advertising tailored to a person’s location, part of a more than $20 billion market for location-based ads last year.

{ NY Times | Continue reading }

GPT-2 is trained with a simple objective: predict the next word, given all of the previous words within some text. […] GPT-2 displays a broad set of capabilities, including the ability to generate conditional synthetic text samples of unprecedented quality, where we prime the model with an input and have it generate a lengthy continuation. […]

SYSTEM PROMPT (HUMAN-WRITTEN)
In a shocking finding, scientist discovered a herd of unicorns living in a remote, previously unexplored valley, in the Andes Mountains. Even more surprising to the researchers was the fact that the unicorns spoke perfect English.

MODEL COMPLETION (MACHINE-WRITTEN, 10 TRIES)
The scientist named the population, after their distinctive horn, Ovid’s Unicorn. These four-horned, silver-white unicorns were previously unknown to science. Now, after almost two centuries, the mystery of what sparked this odd phenomenon is finally solved. […] While examining these bizarre creatures the scientists discovered that the creatures also spoke some fairly regular English. Pérez stated, “We can see, for example, that they have a common ‘language,’ something like a dialect or dialectic.” Dr. Pérez believes that the unicorns may have originated in Argentina, where the animals were believed to be descendants of a lost race of people who lived there before the arrival of humans in those parts of South America.While their origins are still unclear, some believe that perhaps the creatures were created when a human and a unicorn met each other in a time before human civilization. According to Pérez, “In South America, such incidents seem to be quite common.”

{ Open AI | Continue reading }

related { The technology behind OpenAI’s fiction-writing, fake-news-spewing AI, explained }

more { Japanese scientists used A.I. to read minds + NONE of these people exist | Thanks Tim }

quote { Who is Descartes’ Evil Genius? }

As an advanced research topic in forensics science, automatic shoe-print identification has been extensively studied in the last two decades, since shoe marks are the clues most frequently left in a crime scene. […] A large variety of handcrafted features have been used for automatic shoe-print identification. These features have shown good performance in limited and controlled scenarios. Unfortunately, they fail when they are dealing with large intra-class variations caused by the noise, oc- clusions, rotation and various scale distortions. A good alternative to these conventional features are the learned ones, e.g. deep learning, which have more generalization ability in more complicated scenarios. To be effective, these models need to be trained on a large amount of data.

{ arXiv | PDF }

AI-generated videos that show a person’s face on another’s body are called “deepfakes.” […]

Airbrushing and Photoshop long ago opened photos to easy manipulation. Now, videos are becoming just as vulnerable to fakes that look deceptively real. Supercharged by powerful and widely available artificial-intelligence software developed by Google, these lifelike “deepfake” videos have quickly multiplied across the Internet, blurring the line between truth and lie. […] A growing number of deepfakes target women far from the public eye, with anonymous users on deepfakes discussion boards and private chats calling them co-workers, classmates and friends. Several users who make videos by request said there’s even a going rate: about $20 per fake. […]

Deepfake creators often compile vast bundles of facial images, called “facesets,” and sex-scene videos of women they call “donor bodies.” Some creators use software to automatically extract a woman’s face from her videos and social-media posts. Others have experimented with voice-cloning software to generate potentially convincing audio. […]

The requester of the video with the woman’s face atop the body with the pink off-the-shoulder top had included 491 photos of her face, many taken from her Facebook account. […] One creator on the discussion board 8chan made an explicit four-minute deepfake featuring the face of a young German blogger who posts videos about makeup; thousands of images of her face had been extracted from a hair tutorial she had recorded in 2014. […]

The victims of deepfakes have few tools to fight back. Legal experts say deepfakes are often too untraceable to investigate and exist in a legal gray area: Built on public photos, they are effectively new creations, meaning they could be protected as free speech. […]

Many of the deepfake tools, built on Google’s artificial-intelligence library, are publicly available and free to use. […] Google representatives said the company takes its ethical responsibility seriously, but that restrictions on its AI tools could end up limiting developers pushing the technology in a positive way. […]

“If a biologist said, ‘Here’s a really cool virus; let’s see what happens when the public gets their hands on it,’ that would not be acceptable. And yet it’s what Silicon Valley does all the time,” he said.

{ Washington Post | Continue reading }

Technical experts and online trackers say they are developing tools that could automatically spot these “deepfakes” by using the software’s skills against it, deploying image-recognition algorithms that could help detect the ways their imagery bends belief.

The Defense Advanced Research Projects Agency, the Pentagon’s high-tech research arm known as DARPA, is funding researchers with hopes of designing an automated system that could identify the kinds of fakes that could be used in propaganda campaigns or political blackmail. Military officials have advertised the contracts — code-named “MediFor,” for “media forensics” — by saying they want “to level the digital imagery playing field, which currently favors the manipulator.”

The photo-verification start-up Truepic checks for manipulations in videos and saves the originals into a digital vault so other viewers — insurance agencies, online shoppers, anti-fraud investigators — can confirm for themselves. […]

However, the rise of fake-spotting has spurred a technical blitz of detection, pursuit and escape, in which digital con artists work to refine and craft evermore deceptive fakes. In some recent pornographic deepfakes, the altered faces appear to blink naturally — a sign that creators have already conquered one of the telltale indicators of early fakes, in which the actors never closed their eyes. […] “The counterattacks have just gotten worse over time, and deepfakes are the accumulation of that,” McGregor said. “It will probably forever be a cat-and-mouse game.”

{ Washington Post | Continue reading }

Active, polymorphic material (“Utility Fog”) can be designed as a conglomeration of 100-micron robotic cells (‘foglets’). Such robots could be built with the techniques of molecular nanotechnology […] The Fog acts as a continuous bridge between actual physical reality and virtual reality.

{ NASA | Continue reading }

photo { Joel Meyerowitz, Times Square, New York City, 1963 }

After 4 hours of training, AlphaZero became the strongest chess entity of the planet with an estimated ELO of around 3,400.

{ AlphaZero vs Stockfish 8 | ELO ratings of chess players }

more { How AlphaZero quickly learns each game [chess, shogi, and Go] to become the strongest player in history for each }

related { The ability to distort reality has taken an exponential leap forward with “deep fake” technology. We survey a broad array of responses. | Previously: Researchers can now detect AI-generated fake videos with a 95% success rate }

{ aversion | panic | Thanks Tim }

related { Dick Stain Donald Trump got zero comments for the Stock Market Drop }

Here, we present a method that estimates socioeconomic characteristics of regions spanning 200 US cities by using 50 million images of street scenes gathered with Google Street View cars.

Using deep learning-based computer vision techniques, we determined the make, model, and year of all motor vehicles encountered in particular neighborhoods.

Data from this census of motor vehicles, which enumerated 22 million automobiles in total (8% of all automobiles in the United States), were used to accurately estimate income, race, education, and voting patterns at the zip code and precinct level.

The resulting associations are surprisingly simple and powerful. For instance, if the number of sedans encountered during a drive through a city is higher than the number of pickup trucks, the city is likely to vote for a Democrat during the next presidential election (88% chance); otherwise, it is likely to vote Republican (82%).

{ PNAS | PDF }

photo { Tod Papageorge }

ML is short for machine learning, referring to computer algorithms that can learn to perform particular tasks on their own by analyzing data. AutoML, in turn, is a machine-learning algorithm that learns to build other machine-learning algorithms.

With it, Google may soon find a way to create A.I. technology that can partly take the humans out of building the A.I. systems that many believe are the future of the technology industry. […]

The tech industry is promising everything from smartphone apps that can recognize faces to cars that can drive on their own. But by some estimates, only 10,000 people worldwide have the education, experience and talent needed to build the complex and sometimes mysterious mathematical algorithms that will drive this new breed of artificial intelligence.

The world’s largest tech businesses, including Google, Facebook and Microsoft, sometimes pay millions of dollars a year to A.I. experts, effectively cornering the market for this hard-to-find talent. The shortage isn’t going away anytime soon, just because mastering these skills takes years of work. […]

Eventually, the Google project will help companies build systems with artificial intelligence even if they don’t have extensive expertise.

{ NY Times | Continue reading }

art { Ellsworth Kelly, Concorde I (state), 1981-82 }

{ IEEE | full story | Thanks Tim! }

Google started testing their cars on public roads back in 2009, long before any regulations were even dreamed of. An examination of the California Vehicle Code indicated there was nothing in there prohibiting testing.

For testing purposes, Google has a trained safety driver sitting behind the wheel, ready to take it at any moment. Any attempt to take the wheel or use the pedals disables the automatic systems and the safety driver is in control. The safety drivers took special driving safety courses and were instructed to take control if they have any doubt about safe operation. For example, if a vehicle is not braking as expected when approaching a cross walk, take the controls immediately, do not wait to see if it will detect the pedestrians and stop.

The safety drivers are accompanied by a second person in the passenger seat. Known as the software operator, this person monitors diagnostic screens showing what the system is perceiving and planning, and tells the safety driver if something appeared to be going wrong. The software operator is also an extra set of eyes on the road from time to time.

Many other developers have taken this approach, and some of the regulations written have coded something similar to it into law.

This style of testing makes sense if you consider how we train teen-agers to drive. We allow them to get behind the wheel with almost no skill at all, and a driving instructor sits in the passenger seat. While not required, professional driving instructors tend to have their own brake pedal, and know how and when to grab the wheel if need be. They let the student learn and make minor mistakes, and correct the major ones.

The law doesn’t require that, of course. After taking a simple written test, a teen is allowed to drive with a learner’s permit as long as almost any licenced adult is in the car with them. While it varies from country to country, we let these young drivers get full solo licences after only a fairly simple written test and a short road test which covers only a tiny fraction of situations we will encounter on the road. They then get their paperwork and become the most dangerous drivers on the road.

In contrast, robocar testing procedures have been much more strict, with more oversight by highly trained supervisors. With regulations, there have been requirements for high insurance bonds and special permits to go even further. Both software systems and teens will make mistakes, but the reality is the teens are more dangerous.

{ Brad Templeton | Continue reading }

related { Will You Need a New License to Operate a Self-Driving Car? }

The Random Darknet Shopper is an automated online shopping bot which we provide with a budget of $100 in Bitcoins per week. Once a week the bot goes on shopping spree in the deep web where it randomly choses and purchases one item and has it mailed to us.

Researchers at the MIT are testing out their version of a system that lets them see and analyze what autonomous robots, including flying drones, are “thinking.” […]

The system is a “spin on conventional virtual reality that’s designed to visualize a robot’s ‘perceptions and understanding of the world,’” Ali-akbar Agha-mohammadi, a post-doctoral associate at MIT’s Aerospace Controls Laboratory, said in a statement.

{ LiveScience | Continue reading | via gettingsome }

image { Lygia Clark }

We evaluated the impact of different presentation methods for evaluating how funny jokes are. We found that the same joke was perceived as significantly funnier when told by a robot than when presented only using text.

{ Dr. Hato | PDF }

Ongoing projects: Adding farting to the joking robots.

{ Dr. Hato | Continue reading }

{ Samantha West The Telemarketer Robot Who Swears She’s Not a Robot | more }

{ Thomas Jackson }

Holograms of human figures are appearing increasingly often in airports as virtual assistants. And they may also be introduced in various commercial activities. […]

The woman was two-dimensional, a projection on a human-shaped glass sheet. […] She is a product by Tensator®, a “queue control and management solutions” brand. Installed in June of last year, an aviation trade publication reported she cost the airport only 26,000 dollars. The avatar runs 24 hours a day and is portable so she can be moved to other areas of the terminal. […] You will find similar holographic announcers or “airport virtual assistants” in Dubai, Washington Dulles, Macau, Istanbul Ataturk and Long Beach, among other locations. […] The next step will be to install more interactive virtual assistants, which might answer basic questions from travellers about things like flight times, gates or rental car locations. Their plan is to provide models with a touch-screen interface next to the avatar rather than Siri-style speech technology. Voice recognition, while available in the more expensive models (roughly 100,000 dollars) isn’t recommended for airports due to the likelihood of interference from background noise. […]

Musion is better known for their less practical work: reviving dead celebrity singers. Their most famous project was the digital resurrection of Tupac Shakur at last year’s Coachella Festival. The company also recreated Frank Sinatra to perform at Simon Cowell’s 50th birthday party. […] Copyright permissions and objections from various estates, in addition to the high costs, have so far prevented “resurrections” from becoming a more widespread trend.

{ Domus | Continue reading }

art { Wayne White }

These new robot-painting machines can wash, apply solvent to remove dirt, rinse and then spray two different paint types. […]

Manually, it takes a team of painters 4.5 hours to do the first coat. The robots do it in 24 minutes with perfect quality. Boeing began using the machine in February. By midsummer, all 777 wings will be painted this way. […]

Half the 777 wing-painting team has been redeployed to other roles, such as programming the machines, painting the airline liveries on the fuselage or working on the sophisticated paint job needed for the 787-9 tails, which have a special smooth aerodynamic finish, Clark said.

{ Seattle Times | Continue reading }

From assembly line robots to ATMs and self-checkout terminals, each year intelligent machines take over more jobs formerly held by humans; and experts predict this trend will not stop anytime soon. […]

“By 2015, robots should be able to assist teachers in the classroom. By 2018, they should be able to teach on their own, and this will cause many teachers to lose their jobs.” […]

The ultimate tool to replace doctors could be the nanorobot, a tiny microscopic-size machine that can whiz through veins replacing aging and damaged cells with new youthful ones. This nanowonder with expected development time of mid-to-late 2030s could eliminate nearly all need for human doctors. […]

Experts estimate by 2035, 50 million jobs will be lost to machines […] and by the end of the century, or possibly much sooner, all jobs will disappear. Some believe the final solution will take the form of a Basic Income Guarantee, made available as a fundamental right for everyone. […] America should create a $25,000 annual stipend for every U.S. adult, Brain says, which would be phased in over two-to-three decades. The payments could be paid for by ending welfare programs, taxing automated systems, adding a consumption tax, allowing ads on currency, and other creative ideas.

{ IEET | Continue reading }

Many are only just getting their heads around the idea of 3D printing but scientists at MIT are already working on an upgrade: 4D printing.

At the TED conference in Los Angeles, architect and computer scientist Skylar Tibbits showed how the process allows objects to self-assemble.

It could be used to install objects in hard-to-reach places such as underground water pipes, he suggested.

It might also herald an age of self-assembling furniture, said experts.

{ BBC | Continue reading }