Google Will Improve Machine Vision

Google engineers have developed a machine vision technique able to bring high power visual recognition to ordinary desktop and even mobile computers. It is claimed that the system is able to recognize 100,000 object types within a picture in a few minutes.

The tech giant seems to have improved the fairly standard method of applying convolutional filters to a photo to pick out objects of interest. This appeared tricky, as the filters must need a sample of at least one per object type, i.e. if you are scanning Facebook for cats you will need a filter which identifies cats. In other words, the technique is limited to a small number of categories, or you will need a huge database.

A report, penned by Mark Segal, Tom Dean, Mark Ruzon, Jonathon Shlens, Sudheendra Vijayanarasimhan and Jay Yagnik, describes the method which is able to speed things up by using hashing. The matter is that locality sensitive hashing looks up the results of each step and instead of applying a mask to the pixels and summing the result, it hashes the pixels and then uses them as a lookup in a table of results. In addition, they also use a rank ordering method indicating which filter is the best match for further evaluation.

The experts point out that the change to the basic algorithm has resulted in a speed up of almost 20,000 times faster. With 100,000 object detectors that require more than a million filters to be applied to multiple resolution scalings of the target image, the set up could recognize an object in less than twenty seconds. It should be noted that the hardware used was a single multi-core machine with 20GB of RAM.

Hacking Is Not That Dangerous

It turned out that Intel subsidiary McAfee has exaggerated the cost of hacking in its fundamental study which was used to form the basis for American government cyber security policy. A study was conducted back in 2009 by the Centre for Strategic and International Studies and claimed that hacking cost the world economy $1 trillion.

Now US President keeps quoting this figure, as well as intelligence officials and members of Congress, in order to press for legislation on cybercrime protection. However, it recently turned out that $1 trillion was a huge exaggeration by McAfee. This was revealed by another study that was carried out by CSIS. The outfit discovered multiple flaws in the methodology of McAfee’s study and made a conclusion that a specific number would be much more difficult to calculate. According to the new research, the proper number could be $100 billion to $500 billion – although it’s a pretty high figure, it is hardly a trillion.

This is not the first time the McAfee’s figure was believed to be overstated – before, two principal Microsoft researchers had the same doubts. Actually, the CSIS explained that the United States might lose as little as $20 billion to $25 billion annually to cybercrime or as much as $100 to $140 billion. Both researches were underwritten by McAfee, which is considered to be one of the largest security technology vendors. In the meantime, the bias of the 2009 report was commented on at the time.

Media reports suggest that the companies can have a hard time trying to figure out what was stolen, while there are plenty of more complex economic issues out there which keep the surveys from being accurate. Now, even though the figures are 1/3 lower than the last survey, experts still have doubts that McAfee has got it right. It is clear what McAfee gets out of it if the figure is high – it is known that the company has helped the Department of Defense design a secure infrastructure and the latter has cited McAfee’s $1 trillion overestimate to argue for the expansion of cybersecurity programs.

Although hackers from abroad are definitely an increasing threat, it looks like the arguments for such expansions were based on McAfee’s earlier study results.

Google Nexus 7: Small Tablet To Beat

Google on Wednesday announced an updated Nexus 7 tablet that boosts specs and features across the board. The low-cost 7-inch tablet, which goes on sale as soon as July 30, furthers Google's attempt at being a relevant hardware company.

Google didn't change much about the Nexus 7's cosmetic design. It is about the same size and shape as before, although Google says it is about 6mm narrower thanks to thinner framing. It is about 1.76 ounces lighter. Google's Hugo Barra said the new tablet is much more comfortable to hold due to the thinner and lighter profile. It will be sold in all black and features the same soft-touch finish that adorned the first version. It is being manufactured for Google by Asus.

The Nexus 7's defining feature is a new screen. It has been dramatically improved from 1280 by 800 pixels to 1920 by 1200 pixels, making it a full HD screen. The pixel density improved from 216ppi to 323ppi, a significant jump. Google says it will be an ideal device on which to watch HD content. The Nexus 7's processor also has been improved, from a dual-core chipset to a 1.5-GHz quad-core Qualcomm Snapdragon S4 Pro processor with 2 GB of RAM for apps. The new chip nearly doubles the processing power and boosts graphics performance by a factor of four.

Google added stereo speakers to the Nexus 7 and uses Fraunhofer's virtual surround sound technology to improve how movies and other video content sounds. The speakers are located at the top and bottom, so they provide a stereo effect when the device is held sideways in landscape mode. The Nexus 7 now has two cameras. The user-facing camera rates 1.2 megapixels and the new rear-facing camera rates 5 megapixels. It does not have a flash, but it can record 1080p HD video.

Connectivity options are solid. It ships with 802.11a/b/g/n Wi-Fi, Bluetooth 4.0 Low Energy, HDMI out, NFC and wireless charging via the Qi standard. In addition to a Wi-Fi-only model, the Nexus 7 comes in an LTE variant. A single SKU will support AT&T, T-Mobile, or Verizon Wireless's LTE networks. (Sorry, Sprint fans). The device will be sold unlocked so it can be used with any of those three networks. The battery is good for nine hours of video playback or 10 hours of Web browsing, according to Google.

The Nexus 7 will come in three models: 16 GB with Wi-Fi, 32 GB with Wi-Fi, and 32 GB with Wi-Fi and LTE. The Wi-Fi models will cost $229 for 16 GB and $269 for 32 GB. They both go on sale beginning July 30. The 32-GB Wi-Fi and LTE model will become available in the coming weeks and is priced at $349.

As good as the new hardware is, Google made it plain that it wants users to think of its latest tablet as an entertainment device. The Nexus 7 ships with Android 4.3, which has several new content-friendly features, such as control over user profiles and new DRM tools. Google also officially launched its Play Store-based gaming service, and support for textbooks within Google Play.

With its high-definition screen and low price, the Google Nexus 7 is a bargain.

Google May Offer Eternal Youth to Its Staff

It seems that Google may one day include longer lives into its employees’ employment packages. Indeed, Silicon Valley employment packages are now getting hard to top, especially after the big companies were allowed to headhunt.Google, for example, believes that a longer life is something its employees will be happy with.

According to Todd Carlisle, director of staffing at the search giant, it could reach a point where the perks of working at an organization include extending employee’s life. He believes that if that were the case then employees would likely never leave and would remain incredibly loyal. The most interesting part is that Google is actually headed in that direction.

For example, late in 2012 Google hired Ray Kurzweil, a big fan of the singularity idea, which predicts that life extension technologies would eventually extend people’s lives indefinitely, thus enabling the humanity to start using more robotic bodies. This may start happening in earnest within this decade, but the singularity would require capabilities of the most powerful computer system in the world. And this is exactly what the search giant is building, so Google’s engineers may probably accept lower salaries for the chance to be among the first to get eternal youth. The private joke inside the company is that this approach has one drawback: in case an engineer stops working for Google they could switch him off and wipe his personality from its hard-drive. Or sell him to the NSA, in the light of the latest events.

Google Invites Users To Send Money Via Gmail

Google has begun sending out invitations to try the new version of Google Wallet that allows users to send money via Gmail. The invitations are being sent to those who previously expressed interest in testing the revised Google Wallet.

The company first announced plans to integrate Google Wallet with Gmail at its developer conference in May and made the service available on a viral basis: The only way to try the combined service was to receive funds via Gmail from someone in the initial pool of testers, who had been granted early access.

Google now appears to be ready to generate interest in its Gmail payment capability by more conventional means, though it continues to support viral user acquisition. Those who elect to accept Google's invitation can allow others to join by sending them Gmail messages with cash attached.

Joining requires a bit more information than usual for consumer online services. You must confirm your identity by supplying your address, your date of birth and the last four digits of your social security number. The U.S. government, like other governments around the world, likes to have a way to trace the movement of currency if necessary, for the enforcement of tax and security laws.

[ Another way to pay for your coffee: Starbucks, Square Partner For Mobile Payments. ]

After you've submitted the required information, it should then take an hour or so before the "$" symbol becomes an option in the Gmail message ribbon.

The mechanics of sending money via Gmail are now well documented on a Google support page. It's simply a matter of hovering over the "+" icon at the bottom of a message composition window and then clicking on the "$" icon.

Assuming you've set up Google Wallet and have funds available, all you have to do is enter an amount, click "Attach" and then send your money-laden message. Sending money is free if you use existing Google Wallet funds or funds stored in a linked bank account. Alternately, you can use a credit or debit card as a source of funds, but that imposes a 2.9% transaction fee, $0.30 minimum.

Recipients can claim sent funds upon receipt once they have established a Google Wallet account. It would not be at all surprising if this requirement led to renewed interest in Google Wallet.

Google Wallet adoption has been slow in part because of the abundance of competitors: AT&T, Verizon and T-Mobile USA are backing their own Isis payment service; Apple would rather promote its Passbook mobile payment management system; PayPal and Square have been making inroads in mobile payments; and a group of some two dozen retailers including Wal-Mart and Target want to control their own mobile payment solution.

Our survey finds that most companies will launch mobile commerce within a year. What's your holdup? Also in the new, all-digital Mobile Commerce Takes Off special issue of InformationWeek: Companies that take PCI responsibilities seriously will find that using a cloud provider and staying compliant can be a major project.

3 Reasons Voice Will Finally Come To The Web

Voice is dead. Or at least the digerati think so. It takes some real digging in Silicon Valley to find the voiceheads, the true believers that voice will have its second coming as a Web application.

Today, most people think of Apple's Siri when you say voice app, but what if you could control all your apps with voice, and also search through spoken conversations and find content as easily as you do in email? At the very fringes of consumer and enterprise social interaction, this vision is already here. This emergent paradigm, known as hypervoice, promises to be a major boon for productivity. The real question is whether it will tip and become the next big shift in the Web.

It's kind of crazy that telephony and the Web are still so separate. Voice on the Web is only about the transport of voice, not voice as rich media content. Voice today is like Web 1.0 when Web content simply mimicked brochures. It's so boring, it hurts.

But what if voice was interactive like hypertext? What if we could search, share and find highlights from our conversations -- just like we do with text? Voice could go from a fringe player to a radical new social object with the potential to alter the way we communicate online.

These ideas may seem wild, but they are certainly not new. The voiceheads are quick to pull up their shirts to compare scars. With so many false starts, why is now the time for voice to become a member in good standing of the Web community? Here are three reasons.

1. Productivity #SOS

Today, voice solves only a space problem -- connecting two people across long distances in real time. But that model doesn't line up with how we work today. We work asynchronously, out of our email inboxes and social media activity streams. Live calls are increasingly disruptive to our workflow. Throwing in the pain of connecting across multiple time zones makes the need for a better way to work more pressing.

Text alone can't save us from this time-stretched, overloaded information stream. We need new tools, badly. Emerging hypervoice apps, where we can go back over our voice conversations and quickly find bits of information we need, will be like giving us perfect recall. Imagine augmented memory without an implant.

2. Viva La WebRTC!

The World Wide Web Consortium (W3C) drafted WebRTC as an API definition to enable browser-to-browser applications for voice calling, video chat and peer-to-peer file sharing without plug-ins. Today it's not trivial to put voice on the Web and make the pieces play nicely together, so it's hard to underestimate the impact that WebRTC (Web Real-Time Communication) will have on the development of future voice applications. Although the standard is still gathering form and adherents (e.g., Microsoft and Apple have not joined the party yet), WebRTC promises to make it whip simple for developers to integrate voice and video into their applications. By lowering technology barriers, new applications are likely to emerge quickly and seemingly out of left field. WebRTC will unleash the developers!

3. Say It: "Behavioral Changes"

People are starting to get comfortable talking to, not just through, their devices. We saw this nascent behavior shift start with Siri, and now it is likely to expand with Google's hotwording. These behavioral shifts are a critical step forward, as we have to get comfortable with voice as an interface. We need to move away from using our mobile devices as a typewriter.

It's critical, and yet ... Behavioral shifts are the hardest friction point to overcome. Social convention and etiquette change far slower than our technology advancements.

And while we are talking about barriers, one of the most pressing to overcome for hypervoice will be the acceptance of recording our voice conversations. As an early adopter, I have about two years of recorded conversations. I had assumed that people would be more off-put at the prospect of being "on the record." What has really surprised me is how little anyone seems to care. By regularly recording my conversations in a format that was simply searchable and shareable not only by me, but by them as well, my colleagues saw it as a boon for their own productivity, too.

So the real question is: Are we ready to trade our privacy for productivity? We have done it before, countless times. But in some ways, voice feels special. It feels like part of our personhood. And to this last point, time will only tell.

Link found between game theory and quantum

While research tends to become very specialized and entire communities of scientists can work on specific topics with only a little overlap between them, physicist Dr Nicolas Brunner and mathematician Professor Noah Linden worked together to uncover a deep and unexpected connection between their two fields of expertise: game theory and quantum physics.

Dr Brunner said: "Once in a while, connections are established between topics which seem, on the face of it, to have nothing in common. Such new links have potential to trigger significant progress and open entirely new avenues for research."

Game theory—which is used today in a wide range of areas such as economics, social sciences, biology and philosophy—gives a mathematical framework for describing a situation of conflict or cooperation between intelligent rational players. The central goal is to predict the outcome of the process. In the early 1950s, John Nash showed that the strategies adopted by the players form an equilibrium point (so-called Nash equilibrium) for which none of the players has any incentive to change strategy.

Quantum mechanics, the theory describing the physics of small objects such as particles and atoms, predicts a vast range of astonishing and often strikingly counter-intuitive phenomena, such as quantum nonlocality. In the 1960s, John Stewart Bell demonstrated that the predictions of quantum mechanics are incompatible with the principle of locality, that is, the fact that an object can be influenced directly only by its immediate surroundings and not by distant events. In particular, when remote observers perform measurements on a pair of entangled quantum particles, such as photons, the results of these measurements are highly correlated. In fact, these correlations are so strong that they cannot be explained by any physical theory respecting the principle of locality. Hence quantum mechanics is a nonlocal theory, and the fact that Nature is nonlocal has been confirmed in numerous experiments.

In a paper published in Nature Communications, Dr Brunner and Professor Linden showed that the two above subjects are in fact deeply connected with the same concepts appearing in both fields. For instance, the physical notion of locality appears naturally in games where players adopt a classical strategy. In fact the principle of locality sets a fundamental limit to the performance achievable by classical players (that is, bound by the rules of classical physics).

Next, by bringing quantum mechanics into the game, the researchers showed that players who can use quantum resources, such as entangled quantum particles, can outperform classical players. That is, quantum players achieve better performance than any classical player ever could.

Dr Brunner said: "Such an advantage could, for instance, be useful in auctions which are well described by the type of games that we considered. Therefore, our work not only opens a bridge between two remote scientific communities, but also opens novel possible applications for quantum technologies."

The Easiest Way in the World to Share Files P2P – and How it Works

There are dozens of ways to share files on the Internet but one site has made it so simple that anyone with a Firefox or Chrome browser can share content in seconds. In fact, Sharefest is so simple there’s barely anything to talk about, but behind the scenes it’s a different matter. TorrentFreak caught up with its creators for the lowdown on their super-simple, drag-and-drop, BitTorrent-like swarming system.

Sharing files is now something that many Internet users do on a daily basis, whether that’s documents, images or video.

There are plenty of ways to distribute those files, such as via email, public file-hosting sites, Dropbox-like operations, BitTorrent…the list goes on.

But what if all of those options – even email – are just too complicated?

Sharefest is without doubt the most simple way to share files available today. Fire up the Sharefest website and drag a file to the middle of the page. A link will appear – paste that to as many people as you like. Wait for the transfer to complete. Done.

Simplicity aside, what’s so neat about Sharefest is the technology going on behind the scenes.

Sharefest operates on a mesh network similar to BitTorrent and all transfers are executed peer-to-peer. Files can be shared with many people at once and the technology can swarm to speed up transfers, just like BitTorrent. It’s Open Source and apart from Chrome or Firefox, no additional software is required.

TorrentFreak caught up with Hadar Weiss from Peer5, the outfit behind the project, for the lowdown. Weiss says that the key is bringing file-sharing to the web.

“We are fascinated with the potential of P2P communication within the browser,” Weiss explains.

“We think that P2P is very powerful and there are great implementations such as BitTorrent. But P2P can reach its full potential only when it reaches the Web. Until then, it’s not really for the mainstream. For example, we like to ask ourselves, why can’t my mother use P2P filesharing? The answer, (we believe) is in a user friendly web technology.”

Sharefest uses WebRTC, a technology that can enable browser-to-browser communications such as file-sharing and video/chat applications without need for a plug-in. The technology is present in the latest Chrome and Firefox browsers.

“People talk a lot about the WebRTC revolution, and they mostly refer to the audiovisual world. We believe the change is bigger and that we will see many web applications that use the WebRTC Data API for things other than audio or video. Many of them would need a ‘BitTorrent-like’ many-to-many system, and that’s what we build.

“Sharefest actually started just as the canonical example for many-to-many sharing that uses our technology. We wanted to Open Source some of the stuff we developed at Peer5, and file-sharing was the coolest use-case,” Weiss says.

So, using a real-world example in which I want to share a 250MB video with half a dozen friends, how does that work behind the scenes? Presumably there’s a BitTorrent-like tracker to coordinate transfers among peers who swarm to share the data and speed up downloads?

“There’s a swarming technique, and mesh networking. Meaning that any peer can be connected to number of different peers and can send and receive data. So if there are six others, you might be connected to all of them, or just one,” says Weiss.

“The tracker decides who is connected to you, depending on many parameters (we call it the matching algorithm). The most trivial is what blocks you have. Two peers with roughly the same blocks are rarely connected. Geography, ISP, available bandwidth are also important factors of course, and we will put even more in the future. The nice part, because it is server based, is that peers discover each other very quickly.”

Interestingly, Weiss says that the Sharefest tracker doesn’t need to know the IP addresses of peers.

“To match two peers, the tracker sends the relevant peers a message through a WebSocket channel. They then start a standard WebRTC handshake – SDPs which describe the local peer capabilities are sent through the server, and they start talking with the STUN server. The STUN server help the two peers create the real P2P connection, overcoming firewalls and NATs,” Weiss adds.

Also, Sharefest peers are able to discriminate in a way similar to BitTorrent peers, taking more data from better performing peers while taking less from others.

Sharefest is a work-in-progress and will have more features added over time. One that we’d really like to see is the ability to resume partially completed uploads. Weiss told us that this feature is not only in the works, but just about to be released. Perfect.

US ISP Can Monitor File-Sharing Traffic

At the moment, there are quite a few BitTorrent traffic monitoring firms and systems out there, but AT&T seems to make a premiere as the first Internet service provider to get a patent under which the company is allowed to monitor file-sharing traffic.

The innovative system is supposed to be used by AT&T to predict monitor network congestion before occurring. However, another possible use of it is to check out what material is very popular with pirates on file-sharing portals.

Recently, AT&T has adopted a “six-strikes” anti-piracy program which means that an Internet subscriber alleged of downloading copyrighted content gets 6 warnings before having their Internet connection terminated.

Industry observers have created a chart which allows to better understand how the new system works. According to it, AT&T will maintain RSS feeds of torrents that it could end up downloading and then search through. Although the names of the files will play a huge role in what the system detects as illegal content, it will still go one step further and dig into the file to make sure it is really what it says it is, because everyone knows there can be just a virus hiding under the latest Hollywood blockbuster name.

Apparently, the ISP’s technology will appeal to rights holders because of its possibilities and that they will be able to “convince” the broadband provider (if necessary) to use it as a method to reduce Internet piracy.

Turn Your RSS Feeds into Torrents

Almost anyone uses Rich Site Summary feeds today, as they keep the user updated by converting the required data into a standard format – the RSS document. Now RSS2Torrent can take these documents and turn them into full-pledged torrent feeds.

You know that reading RSS feeds is like reading e-mail – they keep users updated with the latest events based on a list of their favorite portals or blogs. BitTorrent Inc. admits that the relationship between RSS and P2P had always been a professional one, until RSS2Torrent managed to take it one step further.

A new service offered by the company is very easy to use, even for subscribers who are not avid web-surfers. For example, you can get IMDB’s RSS feed link by right-clicking the RSS logo and choosing “Copy Link Location” in the menu. Then you just go to RSS2Torrent and paste the link in question into the box. Click the “Create” button and the conversion will take just a few seconds. You will get a new link, which can be copied and opened with any BitTorrent application.

For instance, uTorrent users should go to the File menu and choose “Add RSS feed”. There they should paste the provided link, which has a look like “http://rss2torrent” and that’s it. The new content will be displayed if you click the “Feed” button on the left side of the app.

Vuze users should look for “Content Discovery” and “Subscriptions” and click the plus to add the desired feed. Then they should choose “Create New Subscription” and click the RSS tab to paste the content and save it. Users are free to choose any RSS feed they want, of course, while being sure that it is regularly updated and ready to be downloaded when the case.

Thanks to TorrentFreak for the source of the article