As it stands, the world’s fastest supercomputer is China’s Tianhe-2, a computer that runs at 33.86 petaflops. For those wondering how fast that is, basically it is capable of performing 33.86 quadrillion floating point operations per second. However it looks like President Obama does not want the US to play second-fiddle to anyone.
So much so that he has issued an executive order (via Motherboard) in which he says that the US must build their own exascale supercomputer, with the goal of being 30 times faster than China’s Tianhe-2. Dubbed the National Strategic Computing Initiative, ultimately this will be a computer that is capable of mimicking the human brain, which apparently isn’t possible with today’s technology.
Today we announced that the Google Translate app now does real-time visual translation of 20 more languages. So the next time you’re in Prague and can’t read a menu, we’ve got your back. But how are we able to recognize these new languages?
In short: deep neural nets. When the Word Lens team joined Google, we were excited for the opportunity to work with some of the leading researchers in deep learning. Neural nets have gotten a lot of attention in the last few years because they’ve set all kinds of records in image recognition. Five years ago, if you gave a computer an image of a cat or a dog, it had trouble telling which was which. Thanks to convolutional neural networks, not only can computers tell the difference between cats and dogs, they can even recognize different breeds of dogs. Yes, they’re good for more than just trippy art—if you’re translating a foreign menu or sign with the latest version of Google’s Translate app, you’re now using a deep neural net. And the amazing part is it can all work on your phone, without an Internet connection. Here’s how.
Step by step
First, when a camera image comes in, the Google Translate app has to find the letters in the picture. It needs to weed out background objects like trees or cars, and pick up on the words we want translated. It looks at blobs of pixels that have similar color to each other that are also near other similar blobs of pixels. Those are possibly letters, and if they’re near each other, that makes a continuous line we should read.
Second, Translate has to recognize what each letter actually is. This is where deep learning comes in. We use a convolutional neural network, training it on letters and non-letters so it can learn what different letters look like.
But interestingly, if we train just on very “clean”-looking letters, we risk not understanding what real-life letters look like. Letters out in the real world are marred by reflections, dirt, smudges, and all kinds of weirdness. So we built our letter generator to create all kinds of fake “dirt” to convincingly mimic the noisiness of the real world—fake reflections, fake smudges, fake weirdness all around.
Why not just train on real-life photos of letters? Well, it’s tough to find enough examples in all the languages we need, and it’s harder to maintain the fine control over what examples we use when we’re aiming to train a really efficient, compact neural network. So it’s more effective to simulate the dirt.
The third step is to take those recognized letters, and look them up in a dictionary to get translations. Since every previous step could have failed in some way, the dictionary lookup needs to be approximate. That way, if we read an ‘S’ as a ‘5’, we’ll still be able to find the word ‘5uper’.
Finally, we render the translation on top of the original words in the same style as the original. We can do this because we’ve already found and read the letters in the image, so we know exactly where they are. We can look at the colors surrounding the letters and use that to erase the original letters. And then we can draw the translation on top using the original foreground color.
Crunching it down for mobile
Now, if we could do this visual translation in our data centers, it wouldn’t be too hard. But a lot of our users, especially those getting online for the very first time, have slow or intermittent network connections and smartphones starved for computing power. These low-end phones can be about 50 times slower than a good laptop—and a good laptop is already much slower than the data centers that typically run our image recognition systems. So how do we get visual translation on these phones, with no connection to the cloud, translating in real-time as the camera moves around?
We needed to develop a very small neural net, and put severe limits on how much we tried to teach it—in essence, put an upper bound on the density of information it handles. The challenge here was in creating the most effective training data. Since we’re generating our own training data, we put a lot of effort into including just the right data and nothing more. For instance, we want to be able to recognize a letter with a small amount of rotation, but not too much. If we overdo the rotation, the neural network will use too much of its information density on unimportant things. So we put effort into making tools that would give us a fast iteration time and good visualizations. Inside of a few minutes, we can change the algorithms for generating training data, generate it, retrain, and visualize. From there we can look at what kind of letters are failing and why. At one point, we were warping our training data too much, and ‘$’ started to be recognized as ‘S’. We were able to quickly identify that and adjust the warping parameters to fix the problem. It was like trying to paint a picture of letters that you’d see in real life with all their imperfections painted just perfectly.
To achieve real-time, we also heavily optimized and hand-tuned the math operations. That meant using the mobile processor’s SIMD instructions and tuning things like matrix multiplies to fit processing into all levels of cache memory.
In the end, we were able to get our networks to give us significantly better results while running about as fast as our old system—great for translating what you see around you on the fly. Sometimes new technology can seem very abstract, and it’s not always obvious what the applications for things like convolutional neural nets could be. We think breaking down language barriers is one great use.
Welcome to the future. No, really, it’s the future, right here and right now. And not just because we’ve got mobile processors that can calculate Pi to the ten trillionth digit, or because our video games are starting to look more like movies than games. Nope, what makes me feel like I’m living in The Future(TM) more than anything else is how all that pie-in-the-sky Moore’s Law tech gets applied to solving very human problems, like figuring out where the exit is in the Jakarta airport.
Case in point: Google’s Translate app is applying the Word Lens visual translate tool, which lets you point your phone’s camera to a sign or piece of paper and see the text in your native language, to 20 new languages. And also La Bamba. When you’re done trying to get Ritchie Valens out of your head, come back here to marvel at how that stuff actually works.
Google let us see the man behind the curtain in this blog post. It’s a detailed explanation of the computer science behind near-instantaneous translation. Basically Google has to use optical character recognition on rotating video frames with an astonishing degree of accuracy, then use a massive cross-language dictionary to translate the words on the screen, then place them back into the video in a position, size, and style that matches the original text… all while using as little transmitted data as possible.
It’s harder than it seems, because the live translation aspect of the tool needs to be able to recognize small, blurry text almost instantly – Google’s computers are basically doing instant Captcha puzzles several times a second. To “train” the system, engineers actually had to create algorithms to add digital dirt and smudges to letters in order to get it to work with as wide a variety of input as possible. That’s even more difficult than it sounds, because Google also had to make sure not to strain the system with false positives or attempt to translate things the user might not intend, like vertical relative text. It also offloads at least some of the work to the phone’s local hardware in order to optimize the connection.
This is the kind of Star Trek magic that makes you wonder what’s coming next. Check out Google’s Research Blog post for a more complete breakdown.
- Google Research Blog
In a project led by investigators at UC San Francisco , scientists have devised a new strategy to precisely modify human immune-system T cells, using the popular genome-editing system known as CRISPR/Cas9. T cells play important roles in a wide range of diseases, from diabetes to AIDS to cancer, so this achievement provides a path toward CRISPR/Cas9-based therapies for many serious health problems, the scientists say. It also provides a versatile new tool for research on T cell function.
The Death Star weapon is here! Japan fires world’s most powerful laser to produce energy equal to 1,000 times the planet’s power consumption
Japan claims to have fired the most powerful laser ever created.
Researchers in Osaka were able to produce a 2-petawatt – or 2 quadrillion-watt – laser beam using the Laser for Fast Ignition Experiments (LFEX).
This is equivalent to 1,000 times the world’s electricity consumption, causing the laser to be compared to that on the Death Star in Star Wars.
This amazing video shows ten incredibly bright spheres lighting up the daytime sky in the middle of Japanese city Osaka.
The ten lights seem to dance around from one side of the skyline to the other in the video that lasts more than two minutes.
It’s been posted by a number of Japanese TV channels and has now been viewed by nearly 180,000 people wondering what the lights could possibly be.
In the description, it says that the balls are ‘quick moving’ and ‘dance around’.
It also claims that what was spotted was similar to a sighting above London’s Hyde Park in June.
What makes us cheat? HORMONES: People with high levels of certain chemicals ‘are more likely to behave badly’
- Study: Higher levels of two hormones in our bodies, testosterone and cortisol, encourage cheating and other unethical behaviour
- Elevated testosterone levels give a person the courage to cheat
- Rasied levels of cortisol, due to stress, give us more of a reason to cheat
Life on Earth may have ‘hiccuped’ into existence from a primordial soup of chemicals, according to new research.
Scientists have found long chains of molecules can develop the ability to replicate themselves, and so pass on their properties, by acting as a template around which other chemicals form.
The findings suggest all that was needed to kickstart the path towards living cells was the formation of these molecular chains, known as polymers rather than a dramatic spark or event.
- Zion Harvey lost both of his hands and feed at the age of two, after the became infected with gangrene
- Earlier this month, the now 8-year-old became the youngest patient ever to undergo a double-hand transplant
- A team of 40 at the Children’s Hospital of Philadelphia labored for 11 hours to complete the breakthrough procedure
- Zion is now learning to use his new set of hands, exercising them by picking up light objects and making small movements
Imagine you’re traveling in a foreign country and you’re not familiar with the language, how would you go about reading signboards and directions? You could ask the locals or you could take a chance and hope you’re down the right patch. If you’re tech savvy you’ll probably whip out your smartphone and use Google Translate’s visual translation feature which as of today supports 27 languages.
The Google Translate app has been updated today and it brings visual translation support for 20 additional languages bringing the total number of supported languages to 27, it also makes real-time voice translations much more faster and smoother.
Users can now translate to and from English and Bulgarian, Catalan, Croatian, Czech, Danish, Dutch, Filipino, Finnish, Hungarian, Indonesian, Lithuanian, Norwegian, Polish, Romanian, Slovak, Swedish, Turkish and Ukrainian.
One-way translations from English to Hindi and Thai are also possible now, the app is capable of easily translating printed text in all of the aforementioned languages.
Users who are interested in trying out these new languages have to first update the Google Translate app and set “English” along with the language that they would like to translate. A small ~2MB language pack install will be required for each language.
The latest Google Translate update is coming to both Android and iOS and it will be rolling out in the next couple of days.
Yesterday we reported that Windows 10’s update files had reportedly begun to download in the background for some users. This also means that there are some users who have yet to receive the update files despite having reserved their copy. This is to be expected since we reckon Microsoft probably wants to stagger its release.
Now if you can’t wait for it to come to your turn, fret not because it seems that there is way to go about forcing the files to download onto your computer, thus allowing you to upgrade to Windows 10 on the spot. This is thanks to the folks at VentureBeat who have published the step-by-step instructions on their website.
So to start things off, users will want to navigate to C:\Windows\SoftwareDistribution\Download, or wherever that particular folder is saved at. You will then need to delete all the files there which basically gives Windows Update a clean slate. Once that’s done, head on over to your Windows Update.
You will then need to launch your Command Prompt by typing “cmd” in the search box, right click it, and then choose to “Run as administrator”. Once Command Prompt is open, type in “wuauclt.exe /updatenow” but do not press enter yet. Switch back to the Windows Update window, click “Check for updates” and once it begins checking, swap back to Command Prompt and press enter.
It will then proceed to download the necessary files you need to install Windows 10. It should be noted that the update files are around 6GB so it could take you awhile. We have yet to try this method for ourselves, but so far it looks like it works for the majority of users. If you’d like to learn more, hit up VentureBeat’s website where there are additional screenshots.
Now that the official launch day is here, some folks who reserved Windows 10 upgrades for their PCs are getting prompts to install the OS via Microsoft’s upgrade tool.
If, however, you don’t want to wait or would prefer to create installation media for a fresh install, you can now download the Windows 10 media creation tool. This tool will let you grab any of the consumer editions of the OS, both 32- and 64-bit versions, and it will create an ISO for burning to optical disk or a bootable USB flash drive.
Here’s hoping it works. I’m making a USB key right now. I’d hate to find out soon that something happened.
Researchers have discovered a material that could break the record for the highest melting point of any substance.
A team of Brown University engineers found that a combination of hafnium, nitrogen, and carbon, in just the right amounts, could withstand 4,400 kelvins, or around 7,460 degrees Fahrenheit. To give you an idea, that’s two-thirds the temperature of the sun’s surface. The outer core of the Earth can hit 4,300 kelvins, for further mind-blowing reference.
The team figured this out through a series of computer simulations that “[infer] melting points by simulating physical processes at the atomic level.”
China is building the world’s largest radio telescope in Guizhou province, south west China.
Construction for the Five hundred metre Aperture Spherical Telescope, shortened to FAST, started in March 2011. The project is now expect to be completed by September 2016.
The gigantic radio telescope will cost approximately 1.2 billion Yuan (£120 million), making it the biggest astronomy project China has ever had.
It will allow scientists to get weaker radio signals from outer space, even further than our solar system, reported People’s Daily Online.
The EM Drive (Electro Magnetic Drive) uses electromagnetic microwave cavities to directly convert electrical energy to thrust without the need to expel any propellant. First proposed by Satellite Propulsion Research, a research company based in the UK founded by aerospace engineer Roger Shawyer, the EM Drive concept was predictably scorned by much of the mainstream research community for allegedly violating the laws of physics, including the conservation of momentum.
However, NASA Eagleworks – an advanced propulsion research group led by Dr. Harold G. “Sonny” White at the Johnson Space Center (JSC) – investigated the EM Drive and presented encouraging test results in 2014 at the 50th Joint Propulsion Conference.
White proposes that the EM Drive’s thrust is due to virtual particles in the quantum vacuum that behave like propellant ions in magneto-hydrodynamical propulsion systems, extracting “fuel” from the very fabric of space-time and eliminating the need to carry propellant. While a number of scientists criticize White’s theoretical model, others feel that he is at least pointing to the right direction. The NASASpaceFlight website and forums have emerged as unofficial news source and discussion space for all things related to the EM Drive and related breakthrough space propulsion proposals such as the Cannae Drive.
Shawyer has often been dismissed by the research establishment for not having peer-reviewed scientific publications, but White and Tajmar have impeccable credentials that put them beyond cheap dismissal and scorn. Physics is an experimental science, and the fact that the EM Drive works is confirmed in the lab. “This is the first time that someone with a well-equipped lab and a strong background in tracking experimental error has been involved, rather than engineers who may be unconsciously influenced by a desire to see it work,” notes Wired referring to Tajmar’s work.
The concept of an EmDrive engine is relatively simple. It provides thrust to a spacecraft by bouncing microwaves around in a closed container.
Solar energy provides the electricity to power the microwaves, which means that no propellant is needed.
The implications for this could be huge. For instance, current satellites could be half the size they are today without the need to carry fuel.
Humans could also travel further into space, generating their own propulsion on the way.
When London-based Roger Sawyer came up with concept in 2000, the only team that took him seriously was a group of Chinese scientists.
In 2009, the team allegedly produced 720 millinewton (or 72g) of thrust, enough to build a satellite thruster. But still, nobody believed they had achieved this.
From driverless cars to wearable technology, Google seems to have its fingers in many pies – but it almost got involved in a burger business, according to new reports.
The search giant tried to buy a start-up developing a cheeseburger made entirely from plants for around $300 million, in its latest bid to build a meat-free world.
But Google had its deal rejected because Impossible foods, the company behind the burger, wanted more cash, according to The Information.
The main selling point for the Moto X Play is its large battery, a 3,3630 mAh unit that promises up to 30 hours of mixed usage. The handset is also TurboPower-enabled: with an optional fast charger, Motorola claims that one can get 8 hours of usage from only 15 minutes of charging time.
A Qualcomm Snapdragon 615 SoC powers the Moto X Play. This chip features an octa-core CPU running at 1.7GHz and an Adreno 405 GPU clocked at 550 MHz. Motorola backs the SoC with 2GB of RAM. The handset can be had with either 16 or 32GB of storage, and it includes a microSD slot with support for cards up to 128GB. Wireless connectivity is provided by dual-band 802.11n Wi-Fi, Bluetooth 4.0 Low Energy, and 4G LTE. NFC is also onboard. Android 5.1.1 ties this extensive feature list together.
The Play’s rear camera boasts a 21-megapixel sensor with an f/2 lens and a dual-tone LED flash. The snapper can capture video at resolutions up to 1080p at 30 frames per second, and it can also capture slow-motion video (though neither resolution nor frame-rate specs are available.) A five-megapixel selfie shooter sits around front.
The Style features a 5.7″ 2560×1440 screen (520 PPI) in a svelte frame, which works out to a 76% screen-to-bezel ratio. Despite the 5.7″ screen, the handset’s dimensions aren’t enormous—154 mm x 76 mm x 11 mm—and it weighs in at 179 g. The phone is powered by a Snapdragon 808 SoC, featuring six cores clocked at 1.8GHz and accompanied by 3GB of RAM. The Style is offered with 16, 32, or 64GB of onboard storage, and there’s a microSD slot good for cards up to 128GB. Motorola touted the Style’s compatibility with all US carriers’ 4G LTE networks, as well.
Motorola says this phone’s camera is one of the best smartphone cameras in the world. It’s a 21-MP unit with an f/2 lens and a dual-tone LED flash. For moving pictures, there’s 1080p video capture at 60 FPS, or 4K capture at 30 FPS. There’s also a 5MP wide-angle selfie snapper at the front with a flash of its own.
The 3,000-mAh battery gets a new trick: an included “TurboPower” charger, which can pack 10 hours of juice into the phone in only 15 minutes of charging. The factory-installed OS is Android 5.1.1, with only minor Motorola customizations. Speaking of customizations, the Style has a variety of back shells and “flip shell” casees available through Motorola’s Maker custom-phone service, with materials ranging from silicone rubber to more exotic offerings like real leather and wood.
- Obama Wants To See The US Build An Exascale Supercomputer
- Google Research Blog Post Shows How WordLens Live Translation Instantly Replaces Text In Live Video
- Scientists successfully edit human immune-system T cells
- The Death Star weapon is here! Japan fires world’s most powerful laser to produce energy equal to 1,000 times the planet’s power consumption
- What makes us cheat? HORMONES: People with high levels of certain chemicals ‘are more likely to behave badly’
- Introducing Amazon Dash Button: Place it. Press it. Get it
- Life on Earth began with ‘hiccups’ rather than a bang
- Zion Harvey who lost limbs to gangrene gets double-hand transplant
- Google Translate Now Supports Visual Translation For 20 More Languages
- Force Your Computer To Update To Windows 10
- Blogs RO
- Famous Quotes
- Food recipises
- Health – Brain
- IT Hardware