Center for Applied Data Ethics suggests treating AI like a bureaucracy

A recent paper from the Center for Applied Data Ethics (CADE) at the University of San Francisco urges AI practitioners to adopt terms from anthropology when reviewing the performance of large machine learning models. The research suggests using this terminology to interrogate and analyze bureaucracy, states, and power structures in order to critically assess the performance of large machine learning models with the potential to harm people.

“This paper centers power as one of the factors designers need to identify and struggle with, alongside the ongoing conversations about biases in data and code, to understand why algorithmic systems tend to become inaccurate, absurd, harmful, and oppressive. This paper frames the massive algorithmic systems that harm marginalized groups as functionally similar to massive, sprawling administrative states that James Scott describes in Seeing Like a State,” the author wrote.

The paper was authored by CADE fellow Ali Alkhatib, with guidance from director Rachel Thomas and CADE fellows Nana Young and Razvan Amironesei.

The researchers particularly look to the work of James Scott, who has examined hubris in administrative planning and sociotechnical systems. In Europe in the 1800s, for example, timber industry companies began using abridged maps and a field called “scientific forestry” to carry out monoculture planting in grids. While the practice resulted in higher initial yields in some cases, productivity dropped sharply in the second generation, underlining the validity of scientific principles favoring diversity. Like those abridged maps, Alkhatib argues, algorithms can both summarize and transform the world and are an expression of the difference between people’s lived experiences and what bureaucracies see or fail to see.

The paper, titled “To Live in Their Utopia: Why Algorithmic Systems Create Absurd Outcomes,” was recently published and accepted by the ACM Conference on Human Factors in Computing Systems (CHI), which will be held in May.

Recalling Scott’s analysis of states, Alkhatib warns against harms that can result from unhampered AI, including the administrative and computational reordering of society, a weakened civil society, and the rise of an authoritarian state. Alkhatib notes that such algorithms can misread and punish marginalized groups whose experiences do not fit within the confines of data considered to train a model.

People privileged enough to be considered the default by data scientists and who are not directly impacted by algorithmic bias and other harms may see the underrepresentation of race or gender as inconsequential. Data Feminism authors Catherine D’Ignazio and Lauren Klein describe this as “privilege hazard.” As Alkhatib put it, “other people have to recognize that race, gender, their experience of disability, or other dimensions of their lives inextricably affect how they experience the world.”

He also cautions against uncritically accepting AI’s promise of a better world.

“AIs cause so much harm because they exhort us to live in their utopia,” the paper reads. “Framing AI as creating and imposing its own utopia against which people are judged is deliberately suggestive. The intention is to square us as designers and participants in systems against the reality that the world that computer scientists have captured in data is one that surveils, scrutinizes, and excludes the very groups that it most badly misreads. It squares us against the fact that the people we subject these systems to repeatedly endure abuse, harassment, and real violence precisely because they fall outside the paradigmatic model that the state — and now the algorithm — has constructed to describe the world.”

At the same time, Alkhatib warns people not to see AI-driven power shifts as inevitable.

“We can and must more carefully reckon with the parts we play in empowering algorithmic systems to create their own models of the world, in allowing those systems to run roughshod over the people they harm, and in excluding and limiting interrogation of the systems that we participate in building.”

Potential solutions the paper offers include undermining oppressive technologies and following the guidance of Stanford AI Lab researcher Pratyusha Kalluri, who advises asking whether AI shifts power, rather than whether it meets a chosen numeric definition of fair or good. Alkhatib also stresses the importance of individual resistance and refusal to participate in unjust systems to deny them power.

Other recent solutions include a culture change in computer vision and NLP, reduction in scale, and investments to reduce dependence on large datasets that make it virtually impossible to know what data is being used to train deep learning models. Failure to do so, researchers argue, will leave a small group of elite companies to create massive AI models such as OpenAI’s GPT-3 and the trillion-parameter language model Google introduced earlier this month.

The paper’s cross-disciplinary approach is also in line with a diverse body of work AI researchers have produced within the past year. Last month, researchers released the first details of OcéanIA, which treats a scientific project for identifying phytoplankton species as a challenge for machine learning, oceanography, and science. Other researchers have advised a multidisciplinary approach to advancing the fields of deep reinforcement learning and NLP bias assessment.

We’ve also seen analysis of AI that teams sociology and critical race theory, as well as anticolonial AI, which calls for recognizing the historical context associated with colonialism in order to understand which practices to avoid when building AI systems. And VentureBeat has written extensively about the fact that AI ethics is all about power.

Last year, a cohort of well-known members of the algorithmic bias research community created an internal algorithm-auditing framework to close AI accountability gaps within organizations. That work asks organizations to draw lessons from the aerospace, finance, and medical device industries. Coauthors of the paper include Margaret Mitchell and Timnit Gebru, who used to lead the Google AI ethics team together. Since then, Google has fired Gebru and, according to a Google spokesperson, opened an investigation into Mitchell.

With control of the presidency and both houses of Congress in the U.S., Democrats could address a range of tech policy issues in the coming years, from laws regulating the use of facial recognition by businesses, governments, and law enforcement to antitrust actions to rein in Big Tech. However, a 50-50 Senate means Democrats may be forced to consider bipartisan or moderate positions in order to pass legislation.

The Biden administration emphasized support for diversity and distaste for algorithmic bias in a televised ceremony introducing the science and technology team on January 16. Vice President Kamala Harris has also spoken passionately against algorithmic bias and automated discrimination. In the first hours of his administration, President Biden signed an executive order to advance racial equality that instructs the White House Office of Science and Technology Policy (OSTP) to participate in a newly formed working group tasked with disaggregating government data. This initiative is based in part on concerns that an inability to analyze such data impedes efforts to advance equity.


VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative technology and transact.

Our site delivers essential information on data technologies and strategies to guide you as you lead your organizations. We invite you to become a member of our community, to access:

  • up-to-date information on the subjects of interest to you
  • our newsletters
  • gated thought-leader content and discounted access to our prized events, such as Transform
  • networking features, and more

Become a member

Repost: Original Source and Author Link

Tech News

Microsoft will shut down Minecraft Earth in June

Microsoft will be shutting down its augmented-reality version of Minecraft, known as Minecraft Earth, beginning in June, 2021, the company said Tuesday. Microsoft blamed the pandemic as a cause, as well as the lack of freedom of movement that the game required.

At its core, Minecraft Earth was a game based upon exploration and discovery. During the early days of lockdown, my kids and I would wander around the neighborhood, looking for animals, ores, and dungeons to explore. But Microsoft said that the “current global situation” prevented the free movement and collaborative play that allowed Minecraft Earth to survive.

As a result, Minecraft Earth will be turned off in June.  “On June 30, we will discontinue all content and service support for the game,” Microsoft said in a blog post. “This means that we will stop all development, and after that date, you will be unable to download or play Minecraft Earth anymore. On July 1, we will delete any Minecraft Earth player data unrelated to Character Creator and Minecoin entitlements.”

Before then, however, Microsoft will release its final update, beginning today. The final update will release all of the content that Microsoft had planned for the game and had completed, and will remove real-world money transactions entirely. If you have made a purchase in Minecraft Earth, however, you’ll be gifted with a free version of the Bedrock version of Minecraft (which you probably own already). All players with paid “ruby balances” will be granted Minecoins, which you can use on the Minecraft Marketplace to purchase skin and texture packs, maps, and even minigames.

The final update also includes:

  • Drastically reduced ruby costs

  • Reduced time requirements for crafting and smelting

  • Replacement of unused crafting & smelting boosts with radius boosts of the same level

  • A set of Character Creator items granted to players who sign in between January 5 and June 30 

“We’re eternally grateful for the amazing talent in the Minecraft community,” Microsoft wrote. “This was not an easy decision, and we’re doing everything we can to make sure you get the most out of Minecraft Earth before it sunsets.”

Note: When you purchase something after clicking links in our articles, we may earn a small commission. Read our affiliate link policy for more details.

Repost: Original Source and Author Link


HDR support finally lands on YouTube for Xbox One, Xbox Series X

It’s been a long time coming, but it looks like YouTube finally has HDR support on Xbox consoles. It doesn’t matter if you’re using a console from this generation or the last, if you’re streaming YouTube on a TV that supports HDR, you should be able to tap into HDR while you watch. This is something that has apparently been in the works for a number of years, but it’s here now at last.

That’s according to Flat Panels HD, at least, which tested YouTube on both an Xbox One S from the previous generation and an Xbox Series X from the current generation. That means HDR support is very likely available on Xbox One X and Xbox Series S as well, despite the fact that YouTube wasn’t tested on those two platforms.

If you’re using one of those consoles and you want to confirm that HDR support is up and running for you, you’ll need to toggle on the “stats for nerds” while using the app. There you’ll see that HDR is active using Google’s VP9-2 codec, which is a rather important thing to keep in mind as Flat Panels HD notes that support for the AV1 codec isn’t live at this time. There’s no telling if Google plans to support AV1 videos in the Xbox YouTube app at some point in the future, but for now, videos using that codec won’t play in HDR on Xbox consoles.

On Xbox One S, it seems that resolution and framerate are limited to 1440p60 while HDR is active. On Xbox Series X, however, resolution takes a bump up to 4K60, which is precisely what we’d expect given Microsoft’s push for the console to support 4K resolution across the board.

So, if you’ve got an Xbox and an HDR capable display connected to it, you can now take full advantage of your display’s capabilities while watching YouTube. We’ll see if the YouTube app is updated in the future to support the AV1 codec, so stay tuned for more.

Repost: Original Source and Author Link


How to Increase Your Internet Speed

There is nothing more frustrating than your internet moving along at a snail’s pace. Webpages take forever to load, and games and streaming services experience lag. With Zoom meetings becoming a more regular occurrence, any significant slowdown in your internet makes working at home much more difficult, too.

Fortunately, you don’t have to suffer through a slow internet speed anymore, especially with so many tips and tricks available to help improve your Wi-Fi speed. Before you reach for your phone to upgrade your service, give these tips a try.

Reset your router

Resetting your router has a lot of advantages for the average home network. It can help dispel hacking attempts, reset the router’s limited memory to help speed things up, and even apply important updates that your router may have been waiting on.

The key is to reboot your router properly. You don’t want to do a factory reset, which will erase all your settings and force you to start from scratch. Avoid the pinholes and on-router reset buttons; instead, simply unplug your router from all connected devices and then from its power source, fully shutting it down. Wait for about a minute for everything to fully turn off, then plug your router back into your power source, modem, and anything else you need. We also have a full guide on what to know about router resetting for more info.

Manage your Wi-Fi channels

Most routers these days are dual-band, which means they offer connections over both the 2.5Ghz frequency and the 5GHz frequency. Some routers are also tri-band, which just means there’s an additional 5GHz band to spread out connections even more.

These different channels exist so that you can apportion device connections across the spectrum and lower the demands for a single channel. This can help speed up your connections, especially if the 2.5GHz band is getting a little crowded.

As a general rule, the 5Ghz band is shorter range but a bit faster, more suitable for devices that are close to the router. The 2.5GHz band is longer range but a little slower, a better choice for devices in other rooms or mobile devices that move around a lot.

Some routers come with automatic allocation features that can assign devices to different channels based on connection needs and switch them to new channels as circumstances change. That’s great, but most routers still don’t have that service, which means you need to go into your router settings and make sure networks are set up for all the bands on your router, then connect each device individually to the channel that’s best for them. It’s a little work, but it really can make a difference for your speeds.

Manage the devices on your network

Google Fi

The average home router can handle around 250 connected devices in theory — in practice, as more devices pile up, bandwidth struggles to cover network needs and slowdowns happen. At the modern home or office where everyone has multiple mobile devices, and smart devices are common, routers can get overstretched and start creating speed problems.

If a lot of new devices have been connected to your Wi-Fi and you are noticing speed problems, you may want to start limiting connected devices. For newer routers, open up your router app and look for the list of connected devices. For older routers, enter your IP address in the browser and search to find your administrator settings, where you should look for a section that says Manage Devices, Restrict Access, or something similar. While settings can vary, you should look for several important options here:

  • Finding unnecessary devices: If devices don’t need to be on your network or it looks like someone is hijacking your Wi-Fi, you can kick these devices off. You can also look for an option to ban their MAC addresses, and make sure you change your Wi-Fi password when finished. It’s not a perfect solution, but it can help remove unknown devices quickly.
  • Throttle bandwidth: Some routers allow you to choose specific connections to throttle, or slow down their speeds. This way, you can give less bandwidth to devices only needed for simple tasks, and more bandwidth to devices used for gaming and streaming, which cuts down on speed problems.
  • Restrict usage: Routers may also have the ability to restrict hours of use for certain devices, which is a good way to make sure devices don’t hog too much bandwidth during key hours of the day. It’s also a reliable way to manage kids’ online connections for busy families.
  • Voice commands: We’re seeing more and more devices with voice command compatibility, especially for Alexa. Saying, “Alexa, disconnect this device” or “Alexa, turn off Timmy’s Wi-Fi” can be useful commands. Alexa can also help you quickly move between router modes (chatting, gaming, standard, etc.) to prioritize certain activities.

Using Ethernet connections

Thus far, we are assuming that you have to use Wi-Fi connections around your home. However, if your computer (or other device) is close enough to the router, then you should consider using a wired Ethernet connection to the ports on your router. This is a very reliable way of increasing speeds, since a wired connection is free of many of the issues that can cause Wi-Fi slowdowns.

“Flush” your DNS

stacked servers By redjar via Flickr

A DNS (domain name system) record keeps track of all the website addresses you visit to better improve future visits. However, DNS records rely on you and the website both staying on their respective servers. If servers change over time — which happens to many websites — the DNS will actually hurt your connection speed because it doesn’t recognize the new server. That’s why it can be a good idea to, as they say, flush your DNS from time to time.

This process doesn’t take long with the right steps. On Windows 10, simply search for Command prompt in the search box, and open the app. Make sure you are running as administrator! Then type the command line, “ipconfig/flushdns” and select enter. Windows will let you know the DNS has been cleared.

Switch to a faster browser

Does your slowdown primarily happen when you use your browser or open a lot of new tabs? Think about switching to a more minimalistic browser that only loads what is necessary. Browsers like Brave for Mac and Opera excel at this, especially if you are willing to tinker with settings. If you haven’t already tried it, Google Chrome is also known for being one of the fastest major browser options.

Add an extension to manage your cache

Your browser cache stores copies of website content to make it easier to reload sites. As internet users learn, when a cache gets too full it can slow down online performance (cookies, history, and similar saved data can also have an impact). Clearing your cache manually can be a pain, which is why it tends to build up over time. We suggest an easier path: Download a Clear Cache extension that will help you customize your clearing actions and immediately clear with a simple button on your browser taskbar.

Consider using a VPN

Mark Coppock/Digital Trends

If you search whether a VPN (virtual private network) can increase your internet speeds, you’ll probably find a lot of conflicting information. Here’s the deal: Some ISPs (internet service providers) will throttle bandwidth based on certain activities, such as hitting a soft data cap or streaming services like Netflix and YouTube. If you have evidence that your ISP is throttling bandwidth like this, a VPN can help by hiding your activity so the ISP doesn’t have the data it needs to make throttling decisions.

However, in many cases a VPN can actually slow your internet speeds down with its combination of encryption and server-hopping. Only try this as a solution if you are sure your ISP is being shady. Fortunately, many VPNs are free to download and enable for a trial period if you want to test it out.

If necessary, upgrade your Wi-Fi system or bandwidth

Many changes are hitting the router market now. Among them are expanded MU-MIMO features for better individual connections. You can also find Wi-Fi 6 and compatible devices and more. These new features are great for improving speed and performance. The catch is that if your router is a couple of years old, you probably have to buy a new version to get these upgrades. You also need to make sure your new mobile devices are compatible with current tech.

You can find your optimal router setup by upgrading. You might boost your performance with mesh routers, using multiple router points around your home.

Improving your Internet package will also get you increased bandwidth. It’s true that you’ll be paying more, but you’ll get the upside of an automatic upgrade, without the inconvenient test period. Ask your provider about any relevant sales and specials—just remember to check the details before agreeing. Remember, too, that you might not achieve the same speeds advertised. After you install the upgrade, you can perform an Internet speed test for accurate results.

Because of the COVID-19 pandemic, several Internet companies are getting rid of data cap restrictions. Overall, providers are working hard to maximize speeds and meet demand, so it’s worth asking if your provider has undergone these improvements before you accept a higher cost. You just might get an increased speed.

Editors’ Choice

Repost: Original Source and Author Link

Tech News

2020 Corolla vs. 2020 Camry

If you’ve already decided on Toyota for their incredible reliability and safety ratings, then next is choosing which sedan will suit you best. 

The smaller Corolla is a popular choice due to the lower price tag, but the Camry offers a bit more space and luxury for the focused buyer. Both are available as a hybrid option, raising the gas mileage to over 50 miles per gallon. Now that we’ve caught your attention, here’s what else they deliver.

Tech features

The Camry offers a wider range of available tech features than the Corolla, which is to be expected because it’s more expensive. It’s available with an updated version of Toyota’s Entune infotainment system, an in-car Wi-Fi hotspot, and a feature called Remote Connect which allows the driver to remotely lock or unlock the cars, keep track of maintenance, and set parameters for guest (or teen) drivers. Some of these features have trickled down to the Corolla, which is more high-tech than it has ever been, but note they’re not all available on the base model.

Joel Patel/Digital Trends

Both sedans come standard with a 7.0-inch touchscreen for the infotainment system, and the software is finally compatible with Android Auto and Apple CarPlay. Toyota rewards buyers who step up in the trim level hierarchy with an 8.0-inch touchscreen regardless of which nameplate they choose. The standard stereo is a six-speaker setup, but both cars are offered with a nine-speaker system at an extra cost. Navigation is optional, too.

Amazon Alexa compatibility, Bluetooth wireless technology, and SiriusXM radio are available, though note you’ll need to pay a monthly fee to enjoy the latter once the free, three-month trial period ends.

Performance and fuel economy

The Corolla’s entry-level engine is a 1.8-liter four-cylinder tuned to deliver 139 horsepower and 126 pound-feet of torque. SE and XSE trim levels receive a bigger, 2.0-liter four that makes 169 hp and 151 lb-ft. of torque. Front-wheel drive comes standard across the board, only Mazda and Subaru offer all-wheel drive in this segment, but transmission options vary depending on displacement. The smaller engine is offered with either a six-speed manual transmission or a continuously variable transmission (CVT). The bigger engine is CVT-only.

With the 1.8-liter, the Corolla returns up to 31 mpg in the city, 40 mpg on the highway, and 34 mpg in a combined cycle, which are stellar figures. With the 2.0-liter, those figures drop slightly to 30, 38, and 33, respectively.

If that’s not efficient enough, the Corolla is available with a 121-horsepower gasoline-electric hybrid powertrain that consists of a 1.8-liter four-cylinder and an electric motor. Fuel economy skyrockets to 53, 52, and 52, respectively.

Base models of the Corolla and Camry get a 1.8-liter and 2.5-liter four-cylinder engine, respectively.

The larger, pricier Camry comes standard with a 2.5-liter four-cylinder engine rated at 203 hp and 184 lb-ft. of torque, though those figures go up ever so slightly to 206 and 186, respectively, in the XSE model. Buyers seeking more power can select a 3.5-liter V6 with 301 hp and 266 lb-ft. of torque on tap. Front-wheel drive and an eight-speed automatic transmission come standard regardless of how many cylinders are under the hood, but the four-cylinder engine can be paired with a winter-taming all-wheel-drive system at an extra cost.

Fuel economy checks in at 21 mpg in the city, 41 mpg on the highway, and 34 mpg in a combined cycle for the four-cylinder engine. Select the six, and those figures drop to 22, 33, and 26, respectively.

The Camry has a hybrid option, too; it’s a Toyota, after all. Its gasoline-electric drivetrain is made up of a 2.5-liter four-cylinder engine and a small motor. The two power sources come together to deliver 208 horsepower while boosting city, highway, and combined fuel economy figures to 51, 52, and 53, respectively.

Interior and exterior design

Both nameplates suffered from a reputation for extraordinary blandness, so designers worked hard to make the current-generation models more exciting than ever before. They may have gone a bit too far; we find the exterior design is a bit overwrought, but the final call is yours to make. The Camry’s interior is just as extroverted as its exterior; its dashboard is characterized by an S-shaped line that divides the center console in two sections. Stylists took the Corolla in a tamer direction with horizontal lines and a screen propped up on the dashboard.

Both cars seat five, but the difference in exterior dimensions is reflected on the inside.

Using a measuring tape highlights the difference between the compact and midsize segments. The compact Corolla measures 182.3 inches long, 70.1 inches wide, 56.5 inches tall, and it rides on a 106.3-inch wheelbase. Note those figures apply to the sedan detailed here; there’s also a hatchback model. The midsize Camry stretches to 192.7 inches in SE and XSE trims (192.1 inches in other trims), 72.4 inches wide, 56.9 inches tall, and rides on a 111.2-inch wheelbase. Simply put, the longer wheelbase clears up more space for people and gear in the cabin.

Both cars seat five passengers in relative comfort, but the Corolla has 88.6 cubic feet of interior volume, while the Camry has 100.4 cubic feet (99.3 cubic feet with a moon roof). The Camry also has 15.1 cubic feet of trunk space, compared to the Corolla’s 13.1 cubic feet. Oddly, both are considered midsize cars by the EPA.


In addition to basic safety features like anti-lock brakes and traction control, both cars come standard with Toyota Safety Sense P, which bundles a pre-collision system that mitigates accidents, lane departure warning, automatic high beams, and adaptive cruise control. The Corolla offers eight standard airbags, while the Camry has 10.


When it comes to pricing, the Corolla wins the budget award. The entry-level Corolla L starts at $19,600. At the other end of the spectrum, the flagship XSE model enters the market at $25,550. That’s more than a Camry L, which carries a base price of $24,425. The fully-loaded XSE V6 rings in at $35,130, which pushes it into luxury car territory. As a side note, Audi’s most affordable model, the A3, starts at $33,300.

Adding options beyond the base models means spending more, of course. If it’s a hybrid powertrain you’re after, plan on shelling out at least $23,100 for the Corolla and $28,430 for the Camry; electrification remains relatively expensive. Whether improved fuel economy is worth a few thousand extra dollars depends on how you drive, where you drive, and how much you’re currently paying for fuel. It may help to sketch out a few scenarios before springing for either Toyota sedan’s pricier hybrid model.

Key rivals

Toyota Camrys and Toyota Corollas each have an extensive set of worthy contenders, with many other car manufacturers producing similar caliber sedans.

Corolla’s biggest rivals originates from either Asia or Europe. Interestingly enough, no American car manufacturers pose a direct threat to the Corolla. The three prominent car manufacturers in Detroit are willingly absent from this particular battle. That being said, our top rivals include the Honda Civic, the Subaru Impreza, the Volkswagen Jetta, the Mazda3, the Nissan Sentra, the Hyundai Elantra, and the Kia Forte.

The Camry has some differences in terms of competition. It is interesting to see some American competition in the form of the Ford Fusion. The remainder of the competition can be narrowed down to the Honda Accord, the Subaru Legacy, the Volkswagen Passat, the Mazda 6, the Nissan Altima, the Hyundai Sonata, and the Kia Optima.

Corollas and Camrys owe their popularity to the reliability of Toyota vehicles. Toyota truly lives up to the name of being one of the most user-friendly and reliable cars on the market. These Toyota sedans are no exception. Both of these models are encouraged investments that are likely to last longer than your average vehicle. The easiest way to discern which of the two is right for you is to compare features. You must decide which exclusive features align with your driving style and preferences.

Editors’ Choice

Repost: Original Source and Author Link

Tech News

Apple report card: How we grade everything released in 2020

What started out with an earnings warning, stock drop, and store closures ended up being a very successful year for Apple. Nearly every product in Apple’s catalog saw a refresh, a couple of new ones joined the fray, and tens of millions of devices were sold to happy customers.

Here’s our 2020 Apple product report card, where we give each segment of Apple’s business a grade from A to F.  Be sure to tune into the Macworld Podcast where our U.S. and U.K. writers debate everything that came out of Cupertino last year.


New products: iPhone SE, iPhone 12 mini, iPhone 12, iPhone 12 Pro, iPhone 12 Pro Max

iphone 12 pro hero01 Jason Cross/IDG

Michael Simon: First came the iPhone SE, which somehow made the iPhone 8 feel fresh again with a speedy chip, better camera, and lower price tag. Then we got OLED displays and 5G support across the board, better cameras, a new design, and a new charging and accessory system in the iPhone 12, along with a larger Max and a new 5.4-inch mini. Whether you’re comparing them to the iPhone 11 or the best Android phones of the year, the iPhone 12 lineup is among the best ever made and pushed the boundaries of what a smartphone can do. Grade: A

Jason Cross: The iPhone 12 is a great lineup, with good sizes, performance, and features across the line. That 5G is included without ballooning the price is a big feather in Apple’s cap. MagSafe has a lot of potential. And let’s not forget the iPhone SE, a really great budget alternative introduced early in the year. Yeah, I want an always-on display and ProMotion, but there’s no denying that these are stellar products. Grade: A

Roman Loyola: I usually buy a new phone every year, but this year I didn’t. My main reason for upgrading is 5G, and it’s just not widespread enough now to make the investment. That’s not Apple’s fault and that doesn’t mean that Apple’s offerings aren’t great—they are, as Michael and Jason point out. Grade: A


New products: MacBook Air (M1), 13-inch MacBook Pro (M1), Mac mini (M1), MacBook Air (Intel), 13-inch MacBook Pro (Intel), 27-inch iMac (Intel)

13in m1 mbp desk angle IDG

Michael Simon: Until November, the Mac was languishing through another year of perfectly fine and utterly forgettable products. But then the M1 landed. Apple shipped new models with the M1 chip and they were faster than nearly every Mac Apple has ever made while keeping the same prices (and even lowering it in the case of the mini). There’s plenty more I want to see—most of all a better webcam in the notebooks—but I’ve never been more excited for the Mac. Grade: A

Repost: Original Source and Author Link


‘Fortnite’ gets soccer skins from major clubs and a Pelé emote

American football has been part of the Fortnite experience for years. But what about the other kind of football? The one that trades sacks and touchdowns for gegenpressing and panenka penalties? Well, the game already has skins inspired by the FIFA World Cup. On January 23rd, however the ludicrously-successful battle royale will be embracing club soccer, too. That means 10 new outfits, with male and female variants, that are compatible with 23 different teams. (So if you own one outfit, you can freely swap between clubs.) The exact outfits are a mystery, but screenshots show a mixture of jerseys and long-sleeve tracksuit tops.

Here’s the full list of compatible teams: Manchester City, Juventus, AC Milan, Inter Milan, AS Roma, Seattle Sounders FC, Atlanta United, Los Angeles FC, Santos FC, Wolverhampton Wanderers, West Ham United, Sevilla, Sporting CP, Borussia Mönchengladbach, FC Schalke 04, VFL Wolfsburg, Rangers, Celtic, Cerezo Osaka, Melbourne City FC, Sydney FC, Western Sydney Wanderers and EC Bahia. It’s a good mixture that spans the Premier League (England), La Liga (Spain), Serie A (Italy) and Bundesliga (Germany), as well as smaller leagues such as the Scottish Premiership (Scotland), MLS (US) and J1 League (Japan). Notably, Epic Games hasn’t included any teams from Ligue 1 (France), such as PSG.

Repost: Original Source and Author Link


Facebook and NYU use artificial intelligence to make MRI scans four times faster

If you’ve ever had an MRI scan before, you’ll know how unsettling the experience can be. You’re placed in a claustrophobia-inducing tube and asked to stay completely still for up to an hour while unseen hardware whirs, creaks, and thumps around you like a medical poltergeist. New research, though, suggests AI can help with this predicament by making MRI scans four times faster, getting patients in and out of the tube quicker.

The work is a collaborative project called fastMRI between Facebook’s AI research team (FAIR) and radiologists at NYU Langone Health. Together, the scientists trained a machine learning model on pairs of low-resolution and high-resolution MRI scans, using this model to “predict” what final MRI scans look like from just a quarter of the usual input data. That means scans can be done faster, meaning less hassle for patients and quicker diagnoses.

“It’s a major stepping stone to incorporating AI into medical imaging,” Nafissa Yakubova, a visiting biomedical AI researcher at FAIR who worked on the project, tells The Verge.

The reason artificial intelligence can be used to produce the same scans from less data is that the neural network has essentially learned an abstract idea of what a medical scan looks like by examining the training data. It then uses this to make a prediction about the final output. Think of it like an architect who’s designed lots of banks over the years. They have an abstract idea of what a bank looks like, and so they can create a final blueprint faster.

“The neural net knows about the overall structure of the medical image,” Dan Sodickson, professor of radiology at NYU Langone Health, tells The Verge. “In some ways what we’re doing is filling in what is unique about this particular patient’s [scan] based on the data.”

The AI software can be incorporated into existing MRI scanners with minimal hassle, say researchers.
Image: FAIR / NYU

The fastMRI team has been working on this problem for years, but today, they are publishing a clinical study in the American Journal of Roentgenology, which they say proves the trustworthiness of their method. The study asked radiologists to make diagnoses based on both traditional MRI scans and AI-enhanced scans of patients’ knees. The study reports that when faced with both traditional and AI scans, doctors made the exact same assessments.

“The key word here on which trust can be based is interchangeability,” says Sodickson. “We’re not looking at some quantitative metric based on image quality. We’re saying that radiologists make the same diagnoses. They find the same problems. They miss nothing.”

This concept is extremely important. Although machine learning models are frequently used to create high-resolution data from low-resolution input, this process can often introduce errors. For example, AI can be used to upscale low-resolution imagery from old video games, but humans have to check the output to make sure it matches the input. And the idea of AI “imagining” an incorrect MRI scan is obviously worrying.

The fastMRI team, though, says this isn’t an issue with their method. For a start, the input data used to create the AI scans completely covers the target area of the body. The machine learning model isn’t guessing what a final scan looks like from just a few puzzle pieces. It has all the pieces it needs, just at a lower resolution. Secondly, the scientists created a check system for the neural network based on the physics of MRI scans. That means at regular intervals during the creation of a scan, the AI system checks that its output data matches what is physically possible for an MRI machine to produce.

A traditional MRI scan created from normal input data, known as k-space data.

An AI-enhanced MRI scan created from a quarter of normal input data.

“We don’t just allow the network to create any arbitrary image,” says Sodickson. “We require that any image generated through the process must have been physically realizable as an MRI image. We’re limiting the search space, in a way, making sure that everything is consistent with MRI physics.”

Yakubova says it was this particular insight, which only came about after long discussions between the radiologists and the AI engineers, that enabled the project’s success. “Complementary expertise is key to creating solutions like this,” she says.

The next step, though, is getting the technology into hospitals where it can actually help patients. The fastMRI team is confident this can happen fairly quickly, perhaps in just a matter of years. The training data and model they’ve created are completely open access and can be incorporated into existing MRI scanners without new hardware. And Sodickson says the researchers are already in talks with the companies that produce these scanners.

Karin Shmueli, who heads the MRI research team at University College London and was not involved with this research, told The Verge this would be a key step to move forward.

“The bottleneck in taking something from research into the clinic, is often adoption and implementation by manufacturers,” says Shmueli. She added that work like fastMRI was part of a wider trend incorporating artificial intelligence into medical imaging that was extremely promising. “AI is definitely going to be more in use in the future,” she says.

Repost: Original Source and Author Link


Apple loses early challenge in Corellium copyright lawsuit

Apple has lost an early challenge in its lawsuit against Corellium, a security firm that offers a virtualized version of iOS for security testing.

Ruling on a motion for summary judgment in the US District Court for Southern Florida, US District Judge Rodney Smith has dismissed Apple’s copyright infringement claim over Corellium’s software, finding that Corellium’s use of Apple code constituted fair use. Smith deferred ruling on a separate Digital Millennium Copyright Act (DMCA) charge, but the result is still a significant setback for the iPhone manufacturer’s legal case. The news was first reported by The Washington Post.

In particular, the judge found that additional features within Corellium’s tool strengthened its case for fair use, particularly the ability to modify the kernel or observe and halt processes.

“Corellium makes several changes to iOS and incorporates its own code to create a product that serves a transformative purpose,” Smith wrote in his ruling. “Hence, Corellium’s profit motivation does not undermine its fair use defense, particularly considering the public benefit of the product.”

Crucially, the court didn’t dismiss all of Apple’s case. Apple has alleged that Corellium circumvented its authentication server and secure boot chain, among other measures, violating the DMCA’s ban on circumventing copy protection measures. Corellium also mounted a fair use defense against the DMCA charges, but the judge did not find it compelling enough to dismiss the DMCA allegations before a full trial.

Apple began its lawsuit against Corellium on copyright grounds in August 2019, adding alleged DMCA violations to the case the following January. Throughout the case, Apple has insisted that its purpose is “not to encumber good-faith security research, but to bring an end to Corellium’s unlawful commercialization of Apple’s valuable copyrighted works.” Still, the lawsuit has been alarming to many security firms who rely on Corellium products for iOS analysis.

The Oracle v. Google case has cast a heavy shadow over the trial, and is referenced several times in the ruling as a preeminent example of successful copyright claim over software. However, the judge ultimately found the case had little bearing on the specific allegations made by Apple.

Repost: Original Source and Author Link


Windows Insider Build 21277 includes 64-bit emulator for Arm

Microsoft is finally bringing Windows on ARM closer to parity with traditional X86 processors. The company has released a beta emulator that allows laptops like its Surface Pro X to run 64-bit X86 apps, though at reduced performance.

The release of the emulator theoretically eliminates one of the challenges Arm laptops have faced: compatibility. Most computers run 64-bit versions of popular apps, simply because it can address four times the memory of an older 32-bit app. With the emulator, those Arm notebooks will, eventually and theoretically, run every Windows app that they’d been unable to before, including popular games.

“When we first launched Windows 10 on ARM in late 2017, the long tail of apps customers needed were dominated by 32-bit-only x86 applications, so we focused our efforts on building an x86 emulator that could run the broad ecosystem of Windows apps seamlessly and transparently,” Hari Pulapaka, Microsoft partner group program manager, wrote in a blog post. “Over time, the ecosystem has moved more toward 64-bit-only x64 apps and we’ve heard the feedback that customers would like to see those x64 apps running on ARM64.”

The emulation capability is part of Windows Insider Build 21277, which was released Thursday evening. It’s part of the Dev Channel, which means the code may never be deployed to the general public—though, given its importance, its release is highly likely. 

Interestingly, only three Windows on Arm laptops—the Samsung Galaxy Book S, the Lenovo Flex 5G, and Microsoft’s own Surface Pro X—have drivers released by Microsoft to run the beta emulator. The links to those drivers are in the blog post

Installing the emulator requires some tweaking, above and beyond simply belonging to the Windows Insider program. A blog post describing the process noted that users will first have to install a preview version of the graphics driver used by the Qualcomm Snapdragon family and the Adreno graphics cores that power Windows on Arm. After that, users will have to install a preview version of the ARM64 C++ redistributable. Fortunately, those steps won’t be required in future builds, Microsoft said.

Unfortunately, even the preview Insider build won’t provide total compatibility. “Some of the apps you try running in emulation may not work initially,” Microsoft warned. Because that emulator has to translate the instructions, performance will be a little slower than it would be with an X86 processor. Microsoft said that both compatibility and performance will improve over time. 

Microsoft’s new build also includes support for Unicode emoji 12.1 and 13.0, and makes a small change to the Start menu’s layout: If a folder in the Start menu contains just a single item, such as an app, it will show just that item instead. But the big news here is that finally, a Windows on Arm laptop can run every app that an X86 chip can. Now, those laptops need to increase performance dramatically to have a chance of competing with Apple’s own M1.

Repost: Original Source and Author Link