×
Robotics

Dyson's Been Secretly Working On Robots That Do Household Chores (engadget.com) 1

Dyson has revealed that it has an entire division that's secretly been developing robot prototypes that do household chores. Engadget reports: The company didn't detail any of the models in particularly, but many look like regular robot arms adapted to do specialized home chores like cleaning and tidying. One appeared to be designed to vacuum out the seat cushions, mapping an armchair out in detail to do the job. "So this means I'll never, ever find crisps around the back of my sofa again?" the company's chief engineer, Jake Dyson, asked a researcher in a video (here).

Another robot was putting away dishes or at least placing them in a drying rack, and another was grasping a teddy bear, presumably picking up after a child. Dyson also showed off a "Perception Lab" that was all about robotic vision systems, detecting its environment and mapping humans with sensors, cameras and thermal imaging systems. Dyson is currently on a recruiting drive, looking for around 700 engineers, which is one reason it finally decided to show off the lab (located at Hullavington Airfield, Wiltshire in the UK) after keeping it under wraps.

United States

NY State is Giving Out Hundreds of Robots as Companions For the Elderly (theverge.com) 40

The state of New York will distribute robot companions to the homes of more than 800 older adults. From a report: The robots are not able to help with physical tasks, but function as more proactive versions of digital assistants like Siri or Alexa -- engaging users in small talk, helping contact love ones, and keeping track of health goals like exercise and medication. The scheme is being organized by the New York State Office for the Aging (NYSOFA), and is intended to help address the growing problem of social isolation among the elderly. An estimated 14 million Americans over the age of 65 currently live alone, and this figure is projected to increase over the next decade as the boomer generation ages. Studies have suggested that long-term loneliness is as damaging to an individual's health as smoking.
China

China Is 3D Printing a Massive 590-Foot-Tall Dam, And Constructing It Without Humans (popularmechanics.com) 90

Chinese engineers will take the ideas of a research paper and turn it into the world's largest 3D-printed project. Popular Mechanics: Within two years, officials behind this project want to fully automate the unmanned construction of a 590-foot-tall dam on the Tibetan Plateau to build the Yangqu hydropower plant -- completely with robots. The paper, published last month in the Journal of Tsinghua University (Science and Technology), laid out the plans for the dam, as first reported in the South China Morning Post. Researchers from the State Key Laboratory of Hydroscience and Engineering at Tsinghua University in Beijing explain the backbone of automation for the planned Yellow River dam that will eventually offer nearly five billion kilowatt-hours of electricity annually. (It's worth noting that China's Three Gorges Dam -- a hydroelectric gravity dam spanning the Yangtze River -- is the world's largest power station in terms of energy output.) But it's hard to tell what's more ambitious: the fact that the researchers plan to turn a dam site into effectively a massive 3D-printing project, or that through every step of the process the project eliminates human workers as they go fully robotic.

In the dam-"printing" process, machinery will deliver construction materials to the worksite -- the exact location needed, eliminating human error, they say -- and then unmanned bulldozers, pavers, and rollers will form the dam layer by layer. Sensors on the rollers will keep the artificial intelligence (AI) system informed about the firmness and stability of each of the 3D-printed layers until it reaches 590 feet in height, about the same height as the Shasta Dam in California and shorter than the Hoover Dam's 726 feet. With the largest existing 3D-printed structures rising about 20 feet tall -- from houses in China to an office building in Dubai -- the exploration of 3D-printed projects continues to expand. Already we've seen a 1,640-foot-long retention wall in China, housing and office buildings across the globe, and now the U.S. Army has plans for barracks at Fort Bliss in Texas.

AI

DeepMind Unveils 'Gato' AI Capable of Completing a Wide Range of Complex Tasks (independent.co.uk) 131

An anonymous reader quotes a report from The Independent: Human-level artificial intelligence is close to finally being achieved, according to a lead researcher at Google's DeepMind AI division. Dr Nando de Freitas said "the game is over" in the decades-long quest to realize artificial general intelligence (AGI) after DeepMind unveiled an AI system capable of completing a wide range of complex tasks, from stacking blocks to writing poetry. Described as a "generalist agent," DeepMind's new Gato AI needs to just be scaled up in order to create an AI capable of rivaling human intelligence, Dr de Freitas said.

Responding to an opinion piece written in The Next Web that claimed "humans will never achieve AGI," DeepMind's research director wrote that it was his opinion that such an outcome is an inevitability. "It's all about scale now! The Game is Over!" he wrote on Twitter. "It's all about making these models bigger, safer, compute efficient, faster at sampling, smarter memory, more modalities, innovative data, on/offline... Solving these challenges is what will deliver AGI."

When asked by machine learning researcher Alex Dimikas how far he believed the Gato AI was from passing a real Turing test -- a measure of computer intelligence that requires a human to be unable to distinguish a machine from another human -- Dr de Freitas replied: "Far still." [...] Fielding further questions from AI researchers on Twitter, Dr de Freitas said "safety is of paramount importance" when developing AGI. "It's probably the biggest challenge we face," he wrote. "Everyone should be thinking about it. Lack of enough diversity also worries me a lot."
DeepMind describes Gato in a blog post: "The agent, which we refer to as Gato, works as a multi-modal, multi-task, multi-embodiment generalist policy. The same network with the same weights can play Atari, caption images, chat, stack blocks with a real robot arm and much more, deciding based on its context whether to output text, joint torques, button presses, or other tokens.
Businesses

Uber Launches Robot Food Delivery in California (reuters.com) 24

Uber on Monday said it launched pilot food delivery services with autonomous vehicles in two California cities, and said it was adding electric vehicle charging stations into its global driver app. From a report: The announcements are part of Uber's annual product event where the ride-hail and food delivery company showcases the latest updates to its app. Uber announced one food delivery service using autonomous cars, and a separate pilot using sidewalk robots. Both services are available to Uber Eats users in Santa Monica and West Hollywood in California, and consumers will have the ability to opt out of the programs. The autonomous car pilot is in collaboration with Motional, the self-driving joint venture of Hyundai and Aptiv, and was initially announced in December. read more It launched on Monday, Uber and Motional said. Uber said the sidewalk robots are provided by Serve Robotics, a spin-off of delivery company Postmates, which Uber acquired in 2020.
AI

Swarming Drones Autonomously Navigate a Dense Forest (techcrunch.com) 15

Chinese researchers show off a swarm of drones collectively navigating a dense forest they've never encountered. TechCrunch reports: Researchers at Zheijang University in Hangzhou have succeeded, however, with a 10-strong drone swarm smart enough to fly autonomously through a dense, unfamiliar forest, but small and light enough that each one can easily fit in the palm of your hand. It's a big step toward using swarms like this for things like aerial surveying and disaster response.

Based on an off-the-shelf ultra-compact drone design, the team built a trajectory planner for the group that relies entirely on data from the onboard sensors of the swarm, which they process locally and share with each other. The drones can balance or be directed to pursue various goals, such as maintaining a certain distance from obstacles or each other, or minimizing the total flight time between two points, and so on.

The drones can also, worryingly, be given a task like "follow this human." We've all seen enough movies to know this is how it starts ... but of course it could be useful in rescue or combat circumstances as well. A part of their navigation involves mapping the world around them, of course, and the paper includes some very cool-looking 3D representations of the environments the swarm was sent through. Zhou et alThe study is published in the most recent issue of the journal Science Robotics, which you can read here, along with several videos showing off the drones in action.

Robotics

Autonomous Robots Used In Hundreds of Hospitals At Risk of Remote Hijacks (techcrunch.com) 15

An anonymous reader quotes a report from TechCrunch: [R]esearchers are now finding vulnerabilities in newer hospital technologies that weren't as ubiquitous a decade ago. Enter autonomous hospital robots, the supposed-to-be-friendly self-controlled digital workhorses that can transport medications, bed linens, food, medications and laboratory specimens across a hospital campus. These robots, such as the ones built by robot maker Aethon, are equipped with the space to transport critical goods and security access to enter restricted parts of the hospital and ride elevators, all while cutting labor costs. But researchers at Cynerio, a cybersecurity startup focused on securing hospital and healthcare systems, discovered a set of five never-before-seen vulnerabilities in Aethon robots, which they say allowed malicious hackers to remotely hijack and control these autonomous robots -- and in some cases over the internet.

The five vulnerabilities, which Cynerio collectively call JekyllBot:5, aren't with the robots themselves but with the base servers that are used to communicate with and control the robots that traverse the hallways of the hospitals and hotels. The bugs range from allowing hackers to create new users with high-level access in order to then log in and remotely control the robots and access restricted areas, snoop on patients or guests using the robot's in-built cameras, or otherwise cause mayhem. Asher Brass, the lead researcher on the Aethon vulnerabilities, warned that the flaws required a "very low skill set for exploitation."

Cynerio said the base servers have a web interface that could be accessed from inside the hospital's network, allowing "guest" users to view real-time robot camera feeds and their upcoming schedules and tasks for the day without needing a password. But although the robots' functionality were protected by an "admin" account, the researchers said the vulnerabilities in the web interface could have allowed a hacker to interact with the robots without needing an admin password to log in. One of the five bugs, the researchers said, exposed robots to remote control using a joystick-style controller in the web interface, while exploiting another one of the bugs to interact with door locks, call and ride elevators, and open and close medication drawers.
"The bugs were fixed in a batch of software and firmware updates released by Aethon, after Cynerio alerted the company to the issues," notes TechCrunch. "Aethon is said to have restricted internet-exposed servers to isolate the robots from potential remote attacks, and fixed other web-related vulnerabilities that affected the base station."
Bitcoin

Is Bitcoin Struggling to Find Its Star Power After Miami Conference? (fastcompany.com) 82

Fast Company reviews the Bitcoin 2022 conference held in Miami this week — arguing that it was actually Ethereum that sparked 2021's boom in cryptocurrenies. (And that with NFTs and DAOs, Ethereum still remains the flashier, main driver of popular crypto culture.)

Their conclusion? There's "a real sense of desperation for some kind of star power that can elevate Bitcoin from digital gold for the techno-libertarian set to the true mainstream cultural movement it needs to be in order to actually catch on." In fact, the issue of what influencers or celebrities can do for the Bitcoin community came up directly on Thursday morning, during a panel featuring Odell Beckham Jr., Serena Williams, Aaron Rodgers, and Cash App's crypto product lead, Miles Suter. Beckham and Rodgers have both made headlines recently for taking their salaries in Bitcoin; Williams is heavily involved in the Bitcoin startup world.... It was pretty far away from the high energy radiating from the world of NFTs, and it was clear that the event's bigger names aren't sure what else to do other than just tell the audience to buy Bitcoin over and over again. A lot of people make fun of NFTs, but they're an easier cultural product to point to and talk about than trying to have a fun conversation about lightning networks....

Bitcoin's attempts at going mainstream feel like a real two-steps-forward-one-step-back situation. Its most vocal supporters see it as a war-ending miracle technology. Alex Gladstein, the Human Rights Foundation's chief strategy officer led a panel on Thursday that argued it could lead to peace on Earth. It's not uncommon to hear presenters at Bitcoin 2022 argue that Bitcoin could end all wars forever. But that flies in the face of the conference's wilder, bawdier attractions — the big robot bull statue, the wild after-parties, the endless panels about cancel culture and Twitter drama.

And these competing attitudes within the world of Bitcoin came to a head on Thursday afternoon, when chaos briefly erupted in the conference's main stage when it was announced that Barstool Sports founder Dave Portnoy had dropped out of his much talked-about panel, titled "Bitcoin Is Fuck You Money". Portnoy has not issued any explanation for why he dropped out, but he did spend the rest of the afternoon live tweeting the PGA Masters Tournament. For someone who claims to be all in on the technology, the so-called "baron of Bitcoin" didn't even stay for his panel.

"Fuck you, Dave," the emcee gleefully told the crowd as they cheered and booed in response.

The article reports that other speakers at the conference included:
  • Former presidential candidate Andrew Yang
  • Shark Tank's Kevin O'Leary
  • Paypal and Palantir co-founder Peter Thiel

But the article ultimately argues that the conference "feels like a low-level comic convention that's being held in the same event space as an economic forum."


AI

AI-Powered Artificial Fingertip Gives Robots a Nearly Humanlike Touch (science.org) 7

slashdot reader sciencehabit shares this article from Science magazine: Robots can be programmed to lift a car and even help perform some surgeries, but when it comes to picking up an object they have not touched before, such as an egg, they often fail miserably. Now, engineers have come up with an artificial fingertip that overcomes that limitation. The advance enables machines to sense the textures of these surfaces a lot like a human fingertip does....

[W]hen researchers at the University of Bristol began designing an artificial fingertip in 2009, they used human skin as a guide. Their first fingertip — assembled by hand — was about the size of a soda can. By 2018, they had switched to 3D printing. That made it possible to make the tip and all its components about the size of an adult's big toe and more easily create a series of layers approximating the multilayered structure of human skin. More recently, the scientists have incorporated neural networks into the fingertip, which they call TacTip. The neural networks help a robot quickly process what it's sensing and react accordingly — seemingly just like a real finger.

In our fingertips, a layer of nerve endings deforms when skin contacts an object and tells the brain what's happening. These nerves send "fast" signals to help us avoid dropping something and "slow" signals to convey an object's shape. TacTip's equivalent signals come from an array of pinlike projections underneath a rubbery surface layer that move when the surface is touched. The array's pins are like a hairbrush's bristles: stiff but bendable. Beneath that array is, among other things, a camera that detects when and how the pins move. The amount of bending of the pins provides the slow signal and the speed of bending provides the fast signal. The neural network translates those signals into the fingertip's actions, making it grip more tightly for example, or adjust the angle of the fingertip....

In a second project, Lepora's team added more pins and a microphone to TacTip. The microphone mimics another set of nerve endings deep within our skin that sense vibrations felt as we run our fingers across a surface. These nerve endings enhance our ability to feel how rough a surface is. The microphone did likewise when the researchers tested the enhanced fingertip's ability to differentiate among 13 fabrics.

The article points out that in testing, the artificial fingertip's output "closely matched the neuronal signaling patterns of human fingertips undergoing the same tests."
Robotics

Ai-Da Becomes First Robot To Paint Like An Artist (theguardian.com) 33

Ai-Da is the world's first ultra-realistic humanoid robot that can paint as artists have painted for centuries. The Guardian reports: Devised in Oxford by [Aidan Meller], Ai-Da was created more than two years ago by a team of programmers, roboticists, art experts and psychologists, completed in 2019, and is updated as AI technology improves. She has already demonstrated her ability to sketch and create poems. Her new painting talent was unveiled ahead of the world premier of her solo exhibition at the 2022 Venice Biennale, which opens to the public on 22 April.

Titled Leaping into the Metaverse, Ai-Da Robot's Venice exhibition will explore the interface between human experience and AI technology, from Alan Turing to the metaverse, and will draw on Dante's concepts of purgatory and hell to explore the future of humanity in a world where AI technology continues to encroach on everyday human life. Soon, with the amount of data we freely give about ourselves, and through talking to our phones, computers, cars and even kitchen appliances, AI algorithms "are going to know you better than you do," Meller warned.

We are entering a world, he said, "not understanding which is human and which is machine." "How comfortable are you with that?" "What better thing to have a technological robot artist saying: 'Hang on, are you happy with me doing this?' She is almost daring you to say are you comfortable with this. We are not here to promote robots or technology. We are deeply concerned about the nature of what this technology can do," Meller added. "The whole point of Ai-Da is to highlight what is it we are doing, unknowingly, online all the time."

Biotech

This Snakelike Robot Slithers Down Your Lungs and Could Spot Cancer (msn.com) 28

"Researchers in the United Kingdom have developed an autonomous, snakelike robot designed to slither down human lungs into places that are difficult for medical professionals to reach," reports the Washington Post.

The tool "could improve the detection and treatment of lung cancer and other pulmonary diseases." In a medical paper released in the journal of Soft Robotics last week, scientists from the University of Leeds unveiled a new "magnetic tentacle robot," which is composed of magnetic discs and is roughly 2 millimeters thick — about double the size of a ballpoint pen tip — and less than a-tenth-of-an-inch long.

In the future, the robot's use could be expanded to help doctors better, and more thoroughly, investigate other organs, such as the human heart, kidney or pancreas, they said....

The robot is still 5 to 10 years away from showing up in a clinical setting, researchers said, but the device comes on the heels of a fleet of other robotic innovations allowing doctors the ability to better scan a patient's lungs for cancerous tissue. They are designed to ease a task doctors have long struggled with: reaching the inner recesses of the human body, for diagnostic and treatment purposes, without causing damage or using invasive procedures.... [I]ts smaller size and magnetic composition would allow it to shape-shift more easily and better navigate the intricate shape of a lung's network of airways, which can look like a tree....

Once at its desired location, the robot could ultimately have the capability to take a tissue sample or deliver a clinical treatment.... Nitish V. Thakor, a professor of biomedical engineering at Johns Hopkins University, said the autonomous robot is "very novel and interesting technology" that could become potentially useful in areas outside the lungs, most notably the heart. The device's autonomous capability is its unique factor, he said, and has the capability to change invasive surgeries. "I can imagine a future," he said, "where a full [cancer-screening] CAT scan is done of the lungs, and the surgeon sits down on a computer and lays out this navigation path of this kind of a snake robot and says: 'Go get it.' "

Robotics

Boston Dynamics' Logistics Robot Is Available For Purchase 30

Stretch, a logistics robot from Boston Dynamics that's capable of moving boxes and unloading cargo, is now on sale for anyone who wants to purchase one. Though, as TechCrunch notes, "deliveries are not expected until 2023 and 2024." From the report: The company predictably cites ongoing labor issues as a key driver in interest around the new robot. "Labor shortages and supply chain snags continue to create challenges in keeping the flow of goods moving," says CEO Robert Playter. "Stretch makes logistics operations more efficient and predictable, and it improves safety by taking on one of the most physically demanding jobs in the warehouse. Many of our early adopter customers have already committed to deploying the robot at scale, so we are excited Stretch will soon be put to work more broadly, helping retailers and logistics companies handle the continued surging demand for goods."
Medicine

Walgreens Turns To Robots To Fill Prescriptions (cnbc.com) 66

Walgreens Boots Alliance is opening robot-powered micro-fulfillment centers across the U.S. to fill customers' prescriptions as the role of stores and pharmacists change. CNBC reports: Inside of a large facility in the Dallas area, they fill thousands of prescriptions for customers who take medications to manage or treat high blood pressure, diabetes or other conditions. Each robot can fill 300 prescriptions in an hour, the company said -- roughly the same number that a typical Walgreens pharmacy with a handful of staff may do in a day.

Walgreens Boots Alliance is opening the automated, centralized hubs to keep up in the fast-changing pharmacy industry. The pandemic has intensified the drugstore chain's need to stay relevant as online pharmacies siphon off sales and more customers have items from toilet paper to toothpaste delivered to their doorstep. The global health crisis has also heightened demand for pharmacists, as hospitals and drugstores hired them to administer Covid vaccines and tests. That has forced Walgreens and its competitors, CVS Health and Rite Aid, to rethink the role of their stores and pharmacists.

By 2025, as much as half of Walgreens' prescription volume from stores could be filled at the automated centers, said Rex Swords, who oversees facilities as Walgreens' group president of centralized services, operations and planning. That will free up more of pharmacists' time to provide health care, Brewer said in an interview with CNBC's Bertha Coombs. "We're doing all of this work, so that the pharmacist has an easier job, so that they can get back to being front and center, building a relationship with that patient and interacting the way they were trained -- the work that they love to do," she said. Pharmacists will continue to fill time-sensitive medications and controlled substances at local stores as the company expands its use of robots.

AI

AI Beats Eight World Champions at Bridge (theguardian.com) 20

An artificial intelligence has beaten eight world champions at bridge, a game in which human supremacy has resisted the march of the machines until now. From a report: [...] French startup NukkAI announced the news of its AI's victory on Friday, at the end of a two-day tournament in Paris. The NukkAI challenge required the human champions to play 800 consecutive deals divided into 80 sets of 10. It did not involve the initial bidding component of the game during which players arrive at a contract that they must then meet by playing their cards. Each champion played their own and their "dummy" partner's cards against a pair of opponents. These opponents were the best robot champions in the world to date -- robots that have won many robot competitions but that are universally acknowledged to be nowhere near as good as expert human players.

The AI -- called NooK -- played the same role as the human champion, with the same cards and the same opponents. The score was the difference between those of the human and the AI, averaged over each set. NooK won 67, or 83%, of the 80 sets. Jean-Baptiste Fantun, co-founder of NukkAI, said he had been confident the machine -- which the company has been developing for five years -- would triumph in thousands of deals, but with only 800 it was touch-and-go. Announcing the results, the mathematician Cedric Villani, winner of the Fields medal in 2010, called NukkAI "a superb French success story." AI researcher Veronique Ventos, NukkAI's other co-founder, calls NooK a "new generation AI" because it explains its decisions as it goes along. "In bridge, you can't play if you don't explain," she says.

Math

Linux Random Number Generator Sees Major Improvements (phoronix.com) 80

An anonymous slashdot reader summarizes some important news from the web page of Jason Donenfeld (creator of the open-source VPN protocol WireGuard): The Linux kernel's random number generator has seen its first set of major improvements in over a decade, improving everything from the cryptography to the interface used. Not only does it finally retire SHA-1 in favor of BLAKE2s [in Linux kernel 5.17], but it also at long last unites '/dev/random' and '/dev/urandom' [in the upcoming Linux kernel 5.18], finally ending years of slashdot banter and debate:

The most significant outward-facing change is that /dev/random and /dev/urandom are now exactly the same thing, with no differences between them at all, thanks to their unification in random: block in /dev/urandom. This removes a significant age-old crypto footgun, already accomplished by other operating systems eons ago. [...] The upshot is that every Internet message board disagreement on /dev/random versus /dev/urandom has now been resolved by making everybody simultaneously right! Now, for the first time, these are both the right choice to make, in addition to getrandom(0); they all return the same bytes with the same semantics. There are only right choices.

Phoronix adds: One exciting change to also note is the getrandom() system call may be a hell of a lot faster with the new kernel. The getrandom() call for obtaining random bytes is yielding much faster performance with the latest code in development. Intel's kernel test robot is seeing an 8450% improvement with the stress-ng getrandom() benchmark. Yes, an 8450% improvement.
AI

Self-Driving Trucks Could Replace 90% of Human Long-Distance Truckers, Finds Study (bloombergquint.com) 221

There are already several startups focused on replacing long-haul freight trucks with self-driving trucks, reports Bloomberg — and the potential is huge. (Alternate URLs here and here.) The short trip from a factory or distribution center to an interstate is usually far more complicated than the next several hundred miles. The same is true once the machine exits the interstate. One solution is for trucking companies to set up transfer stations at either end, where human drivers handle the tricky first leg of the trip and then hitch their cargo up to robot rigs for the tiresome middle portion. Another station at the exit would flip the freight back to an analog truck for delivery.

Such a system, according to a new study out of the University of Michigan, could replace about 90% of human driving in U.S. long-haul trucking, the equivalent of roughly 500,000 jobs.

"When we talked to truck drivers, literally every one said, 'Yeah, this part of the job can be automated,'" explained Aniruddh Mohan, a PhD candidate in engineering and public policy at Carnegie Mellon University and a co-author of the study. "We thought they would be a bit more dubious."

There are, however, a handful of big ifs. For one, the autonomous systems would have to figure out how to navigate in crummy weather far better than they can now. Second, regulators in many states still haven't cleared the way for robot rigs. Finally, there's the infrastructure to consider — all the transfer stations where the cargo would pass from the caffeine-fueled analog to the algorithms. Still, if trucking companies focused only on America's Sun Belt, they could fairly easily offset 10% of human driving, the study shows. If they deployed the robots nationwide, but in warmer months only, half of the country's trucking hours could go autonomous.

The article points out that as it is, the workforce of low-paid long-haul truckers "tends to turn over entirely every 12 months or so."

"At the moment, the industry is short about 61,000 drivers, according to the American Trucking Associations."
Robotics

Cornell Researchers Taught a Robot To Take Airbnb Photos (engadget.com) 8

A team of researchers from Cornell University used a computational aesthetic system to teach an AI robot "to not only determine the most pleasing picture in a given dataset, but capture new, original -- and most importantly, good -- shots on its own," writes Engadget's A. Tarantola. The project is called AutoPhoto and was presented last fall at the International Conference on Intelligent Robots and Systems. From the report: This robo-photographer consists of three parts: the image evaluation algorithm, which evaluates a presented image and issues an aesthetic score; a Clearpath Jackal wheeled robot upon which the camera is affixed; and the AutoPhoto algorithm itself, which serves as a sort of firmware, translating the results from the image grading process into drive commands for the physical robot and effectively automating the optimized image capture process.

For its image evaluation algorithm, the Cornell team led by second year Masters student Hadi AlZayer, leveraged an existing learned aesthetic estimation model, which had been trained on a dataset of more than a million human-ranked photographs. AutoPhoto itself was virtually trained on dozens of 3D images of interior room scenes to spot the optimally composed angle before the team attached it to the Jackal. When let loose in a building on campus, as you can see in the video above, the robot starts off with a slew of bad takes, but as the AutoPhoto algorithm gains its bearings, its shot selection steadily improves until the images rival those of local Zillow listings. On average it took about a dozen iterations to optimize each shot and the whole process takes just a few minutes to complete.

"You can essentially take incremental improvements to the current commands," AlZayer told Engadget. "You can do it one step at a time, meaning you can formulate it as a reinforcement learning problem." This way, the algorithm doesn't have to conform to traditional heuristics like the rule of thirds because it already knows what people will like as it was taught to match the look and feel of the shots it takes with the highest-ranked pictures from its training data, AlZayer explained. "The most challenging part was the fact there was no existing baseline number we were trying to improve," AlZayer noted to the Cornell Press. "We had to define the entire process and the problem." Looking ahead, AlZayer hopes to adapt the AutoPhoto system for outdoor use, potentially swapping out the terrestrial Jackal for a UAV. "Simulating high quality realistic outdoor scenes is very hard," AlZayer said, "just because it's harder to perform reconstruction of a controlled scene." To get around that issue, he and his team are currently investigating whether the AutoPhoto model can be trained on video or still images rather than 3D scenes.

Facebook

Facebook Researchers Find Its Apps Can Make Us Lonelier (bloomberg.com) 31

An anonymous reader shares a report: When Facebook hosted an internal competition a few years ago to develop new product ideas, a handful of employees teamed up to build a robot named Max. Shaped like a small, upside-down bowl, Max was designed to be a companion -- a physical device humans could talk to that could detect their mood, according to two people familiar with the hackathon project. The creators gave Max little ears and whiskers so the device would be more fun and approachable, like a cat. Max never evolved beyond the hackathon. But engineers and researchers at the company, now called Meta Platforms, are still grappling with the thorny problem the experimental robot cat was designed to combat: loneliness.

Meta, with a mission to help people connect online, has discovered through internal research that its products can just as easily have an isolating effect. As the company struggles to retain and add users for its already-massive social networks, making sure those people are happy is key to Meta's financial success. Loneliness has come into sharper focus at Meta during the Covid-19 pandemic, as people use its social media apps as alternatives to in-person experiences. Meta has promoted its role as a digital connector, running ads touting its groups and messaging products. "We change the game when we find each other," reads a tagline for one of its recent commercials. But internally, employees are questioning their products' impact on mental health.

Robotics

Lawmakers Express 'Extreme Concern' Over Border Robot Dog Plan (axios.com) 134

A research and development arm of the U.S. Department of Homeland Security announced last month it has been working with the Philadelphia-based company Ghost Robotics to develop a robot dog for the border. Now a small group of Latino U.S. House members recently expressed "extreme concern" about the plan. From a report: A letter obtained by Axios Latino shows that U.S. Reps. Veronica Escobar (D-Texas), Adriano Espaillat (D-NY) and Nanette Diaz Barragan (D-CA) are seeking a meeting with U.S. Customs and Border Protection about the robots. In the letter, the House members write that the term "robot dogs" is a "disingenuous moniker that attempts to soft-pitch the use of this technology."

"It downplays the threat the robots pose to migrants arriving at our southern border and the part they play in a long history of surveillance and privacy violations in our border communities." The letter also said they are concerned that the robot dogs will inevitably result in armed patrols and that they could critically injure, or even kill, migrants or American citizens. Robots used in combination with drones, facial recognition technology and license plate readers, pose civil liberties risks, the letter said.

Robotics

Amazon's Astro Home Robot Remains Elusive Six Months After Debut (bloomberg.com) 15

An anonymous reader shares a report: Last September, Amazon debuted a household robot named Astro that was supposed to usher in -- or at least point to -- a Jetsons-like future. Fifty-three minutes into a press conference otherwise focused on new Ring cameras, a thermostat and a giant Echo speaker with a wall screen, the three-wheeled robot rolled out on stage at the command of Amazon devices chief Dave Limp. With Astro looking on, Limp ticked off the gadget's attributes: advanced computer vision that lets the bot know where it is, home monitoring, media playback and the ability to summon emergency help for elders. Astro would eventually sell for about $1,450, but Limp said people lucky enough to score an invitation could get their hands on one for $1,000 -- or about the price of an iPhone 13 Pro -- and test it out at home.

In a video presentation of the unveiling, Henrik Christensen, a computer science and robotics professor at the University of California at San Diego, said, "Astro is a huge step forward. The next question will be: 'When should I get one?'" A more apt question might have been: When can I get one? Six months later, Astro is tough to find. Hardly anyone is talking about the robot -- which is confounding because early adopters typically love to share their experiences online. A scan for Astro users on YouTube, Twitter and Instagram turned up just two people, who posted brief videos of the bot. Turns out Amazon has so far shipped at most a few hundred Astros, according to people familiar with the situation.

slashdot Top Deals