Uber’s Self-Driving Car Saw the Woman It Killed, Report Says

Uber’s Self-Driving Car Saw the Woman It Killed, Report Says

https://ift.tt/2ILEwX7

The federal investigators examining Uber’s fatal self-driving crash in March released a preliminary report this morning. It lays out the facts of the collision that killed a pedestrian in Tempe, Arizona, and explains what the car actually saw that night.

The National Transportation Safety Board won’t determine the cause of the crash or issue safety recommendations to stop others from happening until it releases its final report, but this first look makes two things clear: Engineering a car that drives itself is very hard. And any self-driving car developer that is currently relying on a human operator to monitor its testing systems—to keep everyone on the road safe—should be extraordinarily careful about the design of that system.

The report says that the Uber car, a modified Volvo XC90 SUV, had been in autonomous mode for 19 minutes and was driving at about 40 mph when it hit 49-year-old Elaine Herzberg as she was walking her bike across the street. The car’s radar and lidar sensors detected Herzberg about six seconds before the crash, first identifying her as an unknown object, then as a vehicle, and then as a bicycle, each time adjusting its expectations for her path of travel.

About a second before impact, the report says, “the self-driving system determined that an emergency braking maneuver was needed to mitigate a collision.” Uber, however, does not allow its system to make emergency braking maneuvers on its own. Rather than risk “erratic vehicle behavior”—like slamming on the brakes or swerving to avoid a plastic bag—Uber relies on its human operator to watch the road and take control when trouble arises.

Furthermore, Uber had turned off the Volvo’s built-in automatic emergency braking system to avoid clashes with its own tech. This is standard practice, experts say. “The vehicle needs one master,” says Raj Rajkumar, an electrical engineer who studies autonomous systems at Carnegie Mellon University. “Having two masters could end up triggering conflicting commands.” But that works a lot better when the master of the moment works the way it’s meant to.

1.3 seconds before hitting Elaine Herzberg, Uber’s car decided emergency braking was necessary—but didn’t have the ability to do that on its own. The yellow bands show distance in meters, and the purple indicates the car’s path.

NTSB

The Robot and the Human

These details of the fatal crash point to at least two serious flaws in Uber’s self-driving system: software that’s not yet ready to replace humans, and humans who were ill-equipped to keep their would-be replacements from doing harm.

Today’s autonomous systems rely on machine learning: They “learn” to classify and respond to situations based on datasets of images and behaviors. The software is shown thousands of images of a cyclist, or a skateboarder, or an ambulance, until it learns to identify those things on its own. The problem is that it’s hard to find images of every sort of situation that could happen in the wild. Can the system distinguish a tumbleweed from a toddler? A unicyclist from a cardboard box? In some of these situations, it should be able to predict the object’s movements, and respond accordingly. In others, the vehicle should ignore the tumbleweed, refrain from a sudden, dangerous braking action, and keep on rolling.

Herzberg, walking a bike loaded with plastic bags and moving perpendicular to the car, outside the crosswalk and in a poorly lit spot, challenged Uber’s system. “This points out that, a) classification is not always accurate, which all of us need to be aware of,” says Rajkumar. “And b) Uber’s testing likely did not have any, or at least not many, images of pedestrians with this profile.”

Solving this problem is a matter of capturing all the strange, unpredictable edge cases on public roads, and figuring out how to train systems to deal with them. It’s the engineering problem at the heart of this industry. It’s supposed to be hard. The car won’t get it right every time, especially not in these early days.

That’s why Uber relied on human safety drivers. And it’s what makes the way they structured their program troubling. At the time, the company’s human operators were paid about $24 an hour (and given plenty of energy drinks and snacks) to work eight-hour shifts behind the wheel. They were told what routes to drive, and what to expect from the software. Above all, they were instructed to keep their eyes on the road at all times, to remain ready to grab the wheel or stomp the brakes. Uber has caught (and fired) drivers who were looking at their phones while on the job—and that shouldn’t surprise anybody.

“We know that drivers, that humans in general, are terrible overseers of highly automated systems,” says Bryan Reimer, a engineer who studies human-machine interaction at MIT. “We’re terrible supervisors. The aviation industry, the nuclear power industry, the rail industry have shown this for decades.”

Yet Uber placed the burden for preventing crashes on the preoccupied shoulders of humans. That’s the tremendous irony here: In its quest to eliminate the humans who cause more than 90 percent of American crashes, which kill about 40,000 people every year, Uber hung its safety system on the ability of a particular human to be perfect.

There other ways to test potentially life-saving tech. Some autonomous developers require two people in every testing vehicle, one to sit behind the wheel and another to take notes on specific events and system failures during the drive. (Uber originally had two operators in each car, but switched to solo drivers late last year.) The safety driver behind the wheel of the crashed Uber told NTSB investigators she wasn’t watching the road in the moments leading up to the collision because she was looking at the car’s interface—which is built into the center console, outside a driver’s natural line of sight. If another human were handling that job, which includes noting observations about the car’s behavior, the person behind the wheel might have spotted Herzberg—and saved her life.

Or, Uber could have given its system the ability to monitor a driver’s attentiveness to the road, and emit a beep or a buzz if it discovers the person behind the wheel isn’t staying on task. Cadillac’s semi-autonomous SuperCruise system uses an infrared camera on the steering column to watch a driver’s head position, and issue warnings when they look away from the road for too long.

Uber’s system didn’t even have a way to alert the driver when it determined emergency braking was necessary, the report says. Many cars on the market today can detect imminent collisions, and alert the driver with flashing red lights or loud beeping. That sort of feature could have helped here. “That is kind of mind-boggling, that the vehicle system did nothing and they had to depend entirely on the driver,” says Steven Shladover, a UC Berkeley research engineer who has spent decades studying automated systems.

Uber says it’s working on its “safety culture,” and has not yet resumed testing, which it paused after the crash. “Over the course of the last two months, we’ve worked closely with the NTSB,” a spokesperson said in a statement. “As their investigation continues, we’ve initiated our own safety review of our self-driving vehicles program.” The company hired former NTSB chair and aviation expert Christopher Hart earlier this month to advise it on safety systems.

Whatever changes Uber makes, they won’t appear in Tempe anytime soon. The company plans to resume testing in Pittsburgh this summer, home to its R&D center. But it’s shutting down its Arizona operation altogether. The move had very human consequences. Uber laid off about 300 workers in the state—many of them safety drivers.

Tech

via Wired Top Stories https://ift.tt/2uc60ci

May 24, 2018 at 02:48PM

Google will always do evil

Google will always do evil

https://ift.tt/2INlc7w


shutterstock

One day in late April or early May, Google removed the phrase “don’t be evil” from its code of conduct. After 18 years as the company’s motto, those three words and chunks of their accompanying corporate clauses were unceremoniously deleted from the record, save for a solitary, uncontextualized mention in the document’s final sentence.

Google didn’t advertise this change. In fact, the code of conduct states it was last updated on April 5th. The “don’t be evil” exorcism clearly took place well after that date.

Google has chosen to actively distance itself from the uncontroversial, totally accepted tenet of not being evil, and it’s doing so in a shady (and therefore completely fitting) way. After nearly two decades of trying to live up to its motto, it looks like Google is ready to face reality.

In order for Google to be Google, it has to do evil.

Exterior view of Google office with Android Marshmallow

This is true for every major technology company. Apple, Facebook, Amazon, Tesla, Microsoft, Sony, Twitter, Samsung, Nintendo, Dell, HP, Toshiba — every one of these organizations can’t compete in the market without engaging in unethical, inhumane and invasive practices. It’s a sliding scale: The larger the company, the more integrated it is in our everyday lives, the more evil it can be.

Take Facebook for example. CEO Mark Zuckerberg will stand onstage at F8 and wax poetic about the beauty of connecting billions of people across the globe, while at the same time patenting technologies to determine users’ social classes and enable discrimination in the lending process, and allowing housing advertisers to exclude racial and ethnic groups, or families with women and children, from their listings.

That’s not even mentioning the Cambridge Analytica scandal and the 85 million Facebook users whose personal information ended up, without permission, in the hands of an overseas political group during the contentious 2016 presidential election.

Mark Zuckerberg on stage at Facebook's F8 Developers Conference 2015

And then there’s Apple, the largest public company in the world. It’s also one of the most secretive, but even so, it’s been caught engaging in evil. Apple is one of the most notorious tech names when it comes to child labor and inhumane working conditions. It’s been tied to child labor in Africa, and the Chinese factories where its phones are assembled are frequently cited over illegal and lethal practices. At least nine workers at Apple’s key factory partner, Foxconn Technology Group, committed suicide in 2010, prompting international outrage. Yet just this year, Bloomberg found iPhone assembly workers in the Catcher Technology Co. factory were required to stand for up to 10 hours a day in heinous conditions, handling chemicals, dealing with loud machines and being exposed to miniscule metal particles without proper masks, gloves, goggles or ear plugs. After their shifts, employees lived in dirty dorms without showers or hot water.

More than 200 workers from a single Samsung production line had died or fallen seriously ill.

Apple isn’t the only tech company to work with Foxconn or Catcher, and it isn’t the only one accused of encouraging inhumane assembly lines. In 2016, the AP reported more than 200 workers from a single Samsung production line had died or fallen seriously ill, many being diagnosed with leukemia, lymphoma and MS, despite being relatively young — in their 20s and early 30s. Samsung has denied any involvement in the lethal trend.

There’s a simple reason major tech companies often look the other way after these scandals, brushing concerns aside as they continue to work with factories known for employing children and operating in barbaric ways. It’s necessity. In order to remain competitive, Apple needs 200 million new iPhones with each updated model, and the most profitable way to make that happen is to partner with Foxconn or Catcher. In Apple’s math, the bottom line outweighs the well-being of workers on the assembly line.

CHINA-SUICIDES

The people who actually work at Apple or any major tech company are not monsters. Ask any Apple employee about child labor in iPhone factories and they’ll assuredly express disgust and outrage — but the company itself is far more powerful than its individualized workforce.

Which brings us back to Google. Earlier this month, roughly a dozen employees quit over the company’s involvement in Project Maven, a military program that aims to use AI systems to analyze drone footage. Though Google insists the technology will be applied to “non-offensive uses only,” some employees are concerned about its potential use in drone strikes. On top of those who quit, nearly 4,000 Google employees have signed a petition demanding the company pull out of Project Maven and refuse to work with the military in the future.

The chances of Google actually cutting ties with the US military are miniscule.

The chances of Google actually cutting ties with the US military are miniscule. Besides, quitting wouldn’t stop Project Maven from moving forward; it would only cut Google out of the process, passing the future of AI drone technology to another company. At least with Google, there’s the underlying promise that these systems won’t be evil.

Well. That was true until just a few weeks ago.

The reason major technology companies have so much power to be evil is because many of them have found ways to do good in our lives. These organizations are big for a reason — Google is the backbone of the internet; Apple is a leader in gadget design and ecosystems; Samsung produces a vast range of devices for a wide swath of people; Facebook truly does connect the world. But as a tech company’s propensity to do good grows, so too does its ability to do terrible things. That’s why Google’s motto — “don’t be evil” — was such a poignant reminder of the humanity necessary to keep these companies in check. Emphasis on the was.

Images: Getty (Google building); pestoverde / Flickr (Mark Zuckerberg); Bobby Yip / Reuters (Foxconn factory)

Tech

via Engadget http://www.engadget.com

May 24, 2018 at 01:36PM

Researchers identify a protein that viruses use as gateway into cells

Researchers identify a protein that viruses use as gateway into cells

https://ift.tt/2s9o1cd

Enlarge /

An electron micrograph of multiple copies of the chikungunya virus.

The word “chikungunya” (chik-en-gun-ye) comes from Kimakonde, the language spoken by the Makonde people in southeast Tanzania and northern Mozambique. It means “to become contorted,” because that’s what happens to people who get infected. The contortion is a result of severe and debilitating joint pain. Chikungunya was first identified in Tanzania in 1952, but by now cases have been reported around the globe. There is no cure; the CDC recommends that “travelers can protect themselves by preventing mosquito bites.”

Chikungunya is only one of a family of viruses transmitted through mosquitoes for which we have no targeted treatment. This may partially be due to the fact that we didn’t know how they get into our cells. But for chikungunya, we’ve just found one of the proteins responsible.

Identification via deletion

Researchers used the CRISPR-Cas9 DNA editing system to delete more than twenty-thousand mouse genes—a different one in each cell in a dish. Then they added chikungunya to the dish, isolated the cells that didn’t get infected, and looked to see which gene they lacked. This gene would encode a protein required for viral infection, since infection didn’t happen in its absence.

In this way they found a gene encoding an adhesion molecule that was required for chikungunya to infect these cells. Similar genes are found in other mammals, birds, and amphibians, and they are homologous to an adhesion molecule used as an entry receptor for another class of viruses. This particular gene goes by the catchy name of Mxra8. Interestingly, no similar protein is found in mosquitoes.

Since the scientists were using a special “cell-culture-adapted vaccine strain” of chikungunya, they repeated their experiment with an Asian strain and a West African strain of the virus. Neither could infect cells lacking Mxra8. Nor could some other viruses in the same family (called arthritogenic alphaviruses): Ross River virus, Mayaro virus, Barmah Forest virus, and O’nyong nyong virus. However, an East/Central/South African strain of chikungunya and a few others in the same family did not seem to be quite as dependent on Mxra8.

In human cells, too

Results were not limited to mouse cells in petri dishes. They also held true in human cells of the various types infected by chikungunya, like fibroblasts, osteoblasts, chondrocytes, and skeletal muscle cells. Humans have four versions of Mxra8, and knocking out each of them diminished the ability of chikungunya to infect the cells. Mice treated with antibodies to Mxra8 had reduced levels of infection—the antibodies bind to the Mxra8 molecules on the surface of the mouse cells, so the virus can’t access it to get in.

Mxra8 doesn’t seem to be required for viral replication, only for viral entry into cells. Further experiments that home in on exactly where the virus binds to it could hopefully lead to the development or identification of small molecules that block the interaction, barring the virus from getting into the cells and preventing infection and disease.

Nature, 2018. DOI: 10.1038/s41586-018-0121-3 (About DOIs).

Tech

via Ars Technica https://arstechnica.com

May 24, 2018 at 12:03PM

Hackers infect 500,000 consumer routers all over the world with malware

Hackers infect 500,000 consumer routers all over the world with malware

https://ift.tt/2IJoCfB

Enlarge /

A Linksys WRVS4400N, one of more than a dozen network devices targeted by VPNFilter.

Hackers, possibly working for an advanced nation, have infected more than 500,000 home and small-office routers around the world with malware that can be used to collect communications, launch attacks on others, and permanently destroy the devices with a single command, researchers at Cisco warned Wednesday.

VPNFilter—as the modular, multi-stage malware has been dubbed—works on consumer-grade routers made by Linksys, MikroTik, Netgear, TP-Link, and on network-attached storage devices from QNAP, Cisco researchers said in an advisory. It’s one of the few pieces of Internet-of-things malware that can survive a reboot. Infections in at least 54 countries have been slowly building since at least 2016, and Cisco researchers have been monitoring them for several months. The attacks drastically ramped up during the past three weeks, including two major assaults on devices located in Ukraine. The spike, combined with the advanced capabilities of the malware, prompted Cisco to release Wednesday’s report before the research is completed.

Expansive platform serving multiple needs

“We assess with high confidence that this malware is used to create an expansive, hard-to-attribute infrastructure that can be used to serve multiple operational needs of the threat actor,” Cisco researcher William Largent wrote. “Since the affected devices are legitimately owned by businesses or individuals, malicious activity conducted from infected devices could be mistakenly attributed to those who were actually victims of the actor. The capabilities built into the various stages and plugins of the malware are extremely versatile and would enable the actor to take advantage of devices in multiple ways.”

Sniffers included with VPNFilter collect login credentials and possibly supervisory control and data acquisition traffic. The malware also makes it possible for the attackers to obfuscate themselves by using the devices as nondescript points for connecting to final targets. The researchers also said they uncovered evidence that at least some of the malware includes a command to permanently disable the device, a capability that would allow the attackers to disable Internet access for hundreds of thousands of people worldwide or in a focused region, depending on a particular objective.

“In most cases, this action is unrecoverable by most victims, requiring technical capabilities, know-how, or tools that no consumer should be expected to have,” Cisco’s report stated. “We are deeply concerned about this capability, and it is one of the driving reasons we have been quietly researching this threat over the past few months.”

Tech

via Ars Technica https://arstechnica.com

May 23, 2018 at 03:18PM

Comcast confirms plan to buy 21st Century Fox and control of Hulu

Comcast confirms plan to buy 21st Century Fox and control of Hulu

https://ift.tt/2s7sdJl

Comcast today said it is preparing an offer to buy major portions of 21st Century Fox, which would give it majority control of Hulu and other media properties.

Walt Disney Company already has a $52.4 billion all-stock deal to buy the 21st Century Fox properties. But Comcast was rumored to be lining up $60 billion in financing in order to make a hostile bid for the Fox assets, and Comcast’s announcement today confirms it.

Comcast “is considering, and is in advanced stages of preparing, an offer for the businesses that Fox has agreed to sell to Disney,” Comcast’s announcement said. Comcast is working on the offer in preparation for shareholder meetings in which the Disney/Fox deal will be considered.

The Fox properties for sale do not include assets such as the Fox News Channel, Fox Business Network, and Fox Broadcasting Company. Those properties would be spun off into a company being referred to as “New Fox,” and Comcast would acquire 21st Century Fox after the spinoff.

The Fox sale to either Disney or Comcast would include 21st Century Fox’s film and television studios; cable entertainment networks; the Fox Sports Regional Networks; and international properties including Star in India and Fox’s 39-percent ownership of Sky across Europe.

Comcast seeks control of Hulu

The sale would also include Fox’s 30-percent stake in Hulu, the popular online video streaming service. Comcast already owns 30 percent of Hulu, so a deal with Fox would give the nation’s largest cable company majority control over the online video provider.

Comcast said its offer for Fox “would be at least as favorable to Fox shareholders as the Disney offer.”

“While no final decision has been made, at this point the work to finance the all-cash offer and make the key regulatory filings is well advanced,” Comcast also said.

Separately, Comcast is trying to buy Sky, the British pay-TV company. After an initial review, a British government official said the government is unlikely to challenge a Comcast/Sky merger.

In the US, a Comcast/Fox deal would face scrutiny from the Department of Justice, which is trying to stop AT&T from buying Time Warner Inc. A judge is expected to rule on the AT&T/Time Warner merger by June 12, and Comcast was reportedly waiting for the outcome of that trial before proceeding with a bid for Fox assets.

“Comcast made the announcement Wednesday about its preparations because its executives were worried that Fox and Disney might rush a shareholder vote before the decision on the AT&T/Time Warner deal came down, according to people close to Comcast,” The Wall Street Journal reported today.

Comcast already owns NBCUniversal thanks to a 2011 acquisition; it was the NBC purchase that gave Comcast its minority stake in Hulu. US regulators imposed merger conditions that prevent Comcast from exercising operational control over Hulu, but those conditions will expire on September 1 of this year.

Tech

via Ars Technica https://arstechnica.com

May 23, 2018 at 01:22PM

Echoes Of Cuba? U.S. Employee In China Hit With ‘Sensations Of Sound And Pressure’

Echoes Of Cuba? U.S. Employee In China Hit With ‘Sensations Of Sound And Pressure’

https://ift.tt/2KQwPLQ

An employee of the U.S. government in Guangzhou, China, has reported mysterious symptoms similar to those experienced by State Department employees in Cuba. Here, the U.S. Consulate in Guangzhou.

U.S. Department of State


hide caption

toggle caption

U.S. Department of State

An employee of the U.S. government in Guangzhou, China, has reported mysterious symptoms similar to those experienced by State Department employees in Cuba. Here, the U.S. Consulate in Guangzhou.

U.S. Department of State

The State Department said that a U.S. government employee assigned to Guangzhou, China, has reported experiencing “vague, but abnormal, sensations of sound and pressure.”

Secretary of State Mike Pompeo said on Wednesday that “the medical indications are very similar and entirely consistent” with the symptoms reported by Americans working at the U.S. Embassy in Cuba. “We have medical teams that are moving to be on the ground there. We are working to figure out what took place both in Havana, and now in China, as well.”

The federal employee reported experiencing physical symptoms from late 2017 through April 2018, when he or she returned to the U.S. for medical evaluation. That evaluation found that the employee’s symptoms were similar to those of someone with a head concussion or mild traumatic brain injury.

The U.S. operates a consulate in Guangzhou, a sprawling commercial city and port in southern China. The Chinese government has assured the U.S. that it is investigating the matter and taking appropriate measures, according to a State Department spokesperson.

Suspected “sonic attacks” affecting more than a dozen U.S. diplomats and family members in Havana beginning in November 2016 led to the U.S. pulling all nonessential staff from the embassy and expelling 15 Cuban diplomats last year. Cuba has denied any involvement in the attacks.

Though some experts expressed doubt when the possibility of a sonic attack in Cuba was first raised, the State Department points to a study by independent medical personnel published earlier this year in The Journal of the American Medical Association.

That study found that 21 U.S. government employees in Havana experienced “persistent cognitive, vestibular, and oculomotor dysfunction, as well as sleep impairment and headaches, were observed … associated with reports of directional audible and/or sensory phenomena of unclear origin. These individuals appeared to have sustained injury to widespread brain networks without an associated history of head trauma.”

Dr. Charles Rosenfarb, the State Department’s medical director, told a Senate subcommittee in January that the employees “associated the onset of these symptoms to their exposures with unusual sounds or auditory sensations. Various descriptions were given: ‘a high-pitched beam of sound’; an ‘incapacitating sound’; a ‘baffling sensation’ akin to driving with the windows partially open in a car; or just an intense pressure in one ear.”

The State Department says it doesn’t know what caused the reported symptoms in Guangzhou, and isn’t aware of any other similar cases in China.

But it did issue a warning Wednesday: “While in China, if you experience any unusual acute auditory or sensory phenomena accompanied by unusual sounds or piercing noises, do not attempt to locate their source. Instead, move to a location where the sounds are not present.”

The U.S. government has about 2,000 employees posted to China. The mysterious incident comes at a time of tension between the U.S. and China over trade disputes.

News

via NPR Topics: News https://ift.tt/2m0CM10

May 23, 2018 at 11:53AM