Home Imrprovement Designs: July 2017

Monday, July 31, 2017

Machines Are Developing Language Skills Inside Virtual Worlds

Machines Are Developing Language Skills Inside Virtual Worlds

It’s hard to teach machines to use language. That’s why they should teach themselves instead.

http://ift.tt/2ueBIpf

An email prankster fooled a number of White House officials, including official tasked with cyber security, by posing as other officials (Jake Tapper/CNN)

Jake Tapper / CNN:
An email prankster fooled a number of White House officials, including official tasked with cyber security, by posing as other officials — (CNN)A self-described “email prankster” in the UK fooled a number of White House officials into thinking he was other officials, including an episode …

http://ift.tt/2hi3Ard

Amazon suspends sales of BLU phones amid renewed concerns that pre-loaded software is sending user data to China (Roger Cheng/CNET)

Roger Cheng / CNET:
Amazon suspends sales of BLU phones amid renewed concerns that pre-loaded software is sending user data to China — Following a report that some Blu phones use an app that collects your data, Amazon has made them unavailable on its site. — Amazon just put budget phone maker Blu in the penalty box.

http://ift.tt/2hiryTb

Latest version of Google's experimental browser Chrome Canary on Android includes setting to block "intrusive" ads (Darrell Etherington/TechCrunch)

Darrell Etherington / TechCrunch:
Latest version of Google's experimental browser Chrome Canary on Android includes setting to block “intrusive” ads — Google will reportedly debut a built-in ad blocking feature for its Chrome browser next year, and now we have the first concrete look at this addition to the web navigation software in action.

http://ift.tt/2f3O7dt

Kleiner Perkins shuts its two-year-old, $4M seed investing program KPCB Edge after all three partners running it leave (Lizette Chapman/Bloomberg)

Lizette Chapman / Bloomberg:
Kleiner Perkins shuts its two-year-old, $4M seed investing program KPCB Edge after all three partners running it leave — Departures of young partners weaken firm's early-stage effort — Kleiner to continue investing broadly in emerging startups — Kleiner Perkins Caufield & Byers shut …

http://ift.tt/2tT5rER

Google says it has begun rolling out Daydream-ready update to Samsung Galaxy S8 and S8+ (Lucas Matney/TechCrunch)

Lucas Matney / TechCrunch:
Google says it has begun rolling out Daydream-ready update to Samsung Galaxy S8 and S8+ — Google's brand of VR is, at last, beginning to arrive on Samsung devices. — Reports emerged last week that certain Verizon Galaxy S8 and S8+ devices were getting the Daydream update, but today Google confirmed the rollout.

http://ift.tt/2uSzptI

Gboard for iPhone adds support for drawing, and sending links to Maps and YouTube videos (Alan Ni/Google)

Alan Ni / Google:
Gboard for iPhone adds support for drawing, and sending links to Maps and YouTube videos — To all the Gboard fans out there: we're adding some clutch additions to your favorite keyboard companion. Now with a tap of the G button, you'll have access to both Maps and YouTube …

http://ift.tt/2uPApR1

Sources: Spotify to double down on original content and fund a new batch of podcasts in coming months, as it promotes shows from established producers (Lucas Shaw/Bloomberg)

Lucas Shaw / Bloomberg:
Sources: Spotify to double down on original content and fund a new batch of podcasts in coming months, as it promotes shows from established producers — Streaming company promotes shows from established producers — Top music site tries to diversify offerings beyond songs

http://ift.tt/2tYcK28

Second Life creator Linden Lab launches open beta for social VR platform Sansar, currently supported on HTC Vive VR headset, the Oculus Rift, and Windows PCs (Dean Takahashi/VentureBeat)

Dean Takahashi / VentureBeat:
Second Life creator Linden Lab launches open beta for social VR platform Sansar, currently supported on HTC Vive VR headset, the Oculus Rift, and Windows PCs — Linden Lab has opened the creator beta version of its Sansar social virtual reality platform to the public.

http://ift.tt/2wfaYX6

Facebook acquired AI startup Ozlo, which had raised $14M, to help Messenger build its personal assistant; Ozlo will shut down its enterprise API and app (Kurt Wagner/Recode)

Kurt Wagner / Recode:
Facebook acquired AI startup Ozlo, which had raised $14M, to help Messenger build its personal assistant; Ozlo will shut down its enterprise API and app — Ozlo is joining Messenger. — Facebook has acquired Ozlo, a Palo Alto-based artificial intelligence startup, to help Messenger build …

http://ift.tt/2vmrAQa

Pandora reports Q2 revenue up 10% YoY to $376.8M, vs. $368.9M est., total subs up 24% YoY to 4.86M, and ticketing service revenue up 31% YoY to $29.7M (Matthew Lynley/TechCrunch)

Matthew Lynley / TechCrunch:
Pandora reports Q2 revenue up 10% YoY to $376.8M, vs. $368.9M est., total subs up 24% YoY to 4.86M, and ticketing service revenue up 31% YoY to $29.7M — Pandora is probably breathing a small sigh of relief after months of chaos have still ended with a positive second quarter, sending the shares up more than 7% in extended trading.

http://ift.tt/2uP9l4t

Tech giants pass on invitation to testify on net neutrality before the House Energy and Commerce Committee; hearing date is pushed back indefinitely (Tony Romm/Recode)

Tony Romm / Recode:
Tech giants pass on invitation to testify on net neutrality before the House Energy and Commerce Committee; hearing date is pushed back indefinitely — They had until today to tell a key House committee, which is extending its deadline. — Amazon, Facebook, Google and Netflix …

http://ift.tt/2tSzBrI

Sources: Spotify passes 60M paying customers, adding over 20M so far in 2017, as it plans for IPO on NYSE later this year (Anna Nicolaou/Financial Times)

Anna Nicolaou / Financial Times:
Sources: Spotify passes 60M paying customers, adding over 20M so far in 2017, as it plans for IPO on NYSE later this year — Unconventional public debut of music streaming service based on subscription success — Read next … Spotify has reached 60m paying customers …

http://ift.tt/2vmaTV2

Reddit raised $200M in funding, is now valued at $1.8B, a16z and Sequoia among investors; sources: Condé Nast owner Advance Publications retains majority stake (Kurt Wagner/Recode)

Kurt Wagner / Recode:
Reddit raised $200M in funding, is now valued at $1.8B, a16z and Sequoia among investors; sources: Condé Nast owner Advance Publications retains majority stake — It's going on a hiring spree and redesigning its website. — Reddit has raised $200 million in new venture funding …

http://ift.tt/2vbKEAe

Hackers breach HBO's network and leak future episodes of Room 104 and Ballers, plus a script allegedly from next week's Game of Thrones episode (James Hibberd/EW.com)

James Hibberd / EW.com:
Hackers breach HBO's network and leak future episodes of Room 104 and Ballers, plus a script allegedly from next week's Game of Thrones episode — HBO chairman: ‘Disruptive, unsettling, and disturbing for all of us’ — HBO has joined the ranks of Hollywood entertainment companies to suffer a major cyber attack.

http://ift.tt/2vYdNfT

Microsoft shuts down its Word Flow keyboard for iOS, suggests downloading SwiftKey instead (Richard Devine/Windows Central)

Richard Devine / Windows Central:
Microsoft shuts down its Word Flow keyboard for iOS, suggests downloading SwiftKey instead — After the experiment has been completed, Microsoft has closed down its Word Flow keyboard on iOS. — Back in February of this year we took a closer look at Microsoft's Word Flow keyboard on iOS …

http://ift.tt/2f1msd7

AMD unveils Radeon RX Vega GPUs for high-end gaming PCs launching August 14, the $499 RX Vega 64 and $399 RX Vega 56 (Ryan Smith/AnandTech)

Ryan Smith / AnandTech:
AMD unveils Radeon RX Vega GPUs for high-end gaming PCs launching August 14, the $499 RX Vega 64 and $399 RX Vega 56 — At this point, one must give credit to AMD for their marketing program for the Radeon RX Vega. The company has opted to dip feed information over many months …

http://ift.tt/2f1jHIW

EPIC files suit with FTC over Google's opaque Store Sales Measurement plan to track shoppers' in-store purchases, saying users have no easy way to opt out (Washington Post)

Washington Post:
EPIC files suit with FTC over Google's opaque Store Sales Measurement plan to track shoppers' in-store purchases, saying users have no easy way to opt out — A prominent privacy rights watchdog is asking the Federal Trade Commission to investigate a new Google advertising program …

http://ift.tt/2tR3Q23

Inside "No More Ransom", a joint law enforcement and private industry effort to fight cybercrime, launched by Europol, Kaspersky, and others last year (Danny Palmer/ZDNet)

Danny Palmer / ZDNet:
Inside “No More Ransom”, a joint law enforcement and private industry effort to fight cybercrime, launched by Europol, Kaspersky, and others last year — No More Ransom launched a year ago: here's the story of how cybersecurity firms and law enforcement are working together to bring down ransomware.

http://ift.tt/2wdHgSg

UK's Callsign raises $35M Series A led by Accel for its AI-based adaptive authentication platform for enterprises (Natasha Lomas/TechCrunch)

Natasha Lomas / TechCrunch:
UK's Callsign raises $35M Series A led by Accel for its AI-based adaptive authentication platform for enterprises — London-based Callsign has closed a $35 million Series A, led by Accel and early stage investor PTB Ventures, for an authentication platform which uses deep learning technology …

http://ift.tt/2vXFJ3p

Snapdeal to end all talks of a sale to Flipkart, and will pursue an independent path; Snapdeal's digital payments unit FreeCharge sold to Axis Bank for $60M (Reuters)

Reuters:
Snapdeal to end all talks of a sale to Flipkart, and will pursue an independent path; Snapdeal's digital payments unit FreeCharge sold to Axis Bank for $60M — Sankalp Phartiyal and Nivedita Bhattacharjee — MUMBAI/BENGALURU (Reuters) - India's Snapdeal said it had decided to remain independent …

http://ift.tt/2tW1ymH

From prototype to product with hybrid neural networks

Apache MXNet and the middle path between declarative and imperative programming.

After several decades as an interest of academic computer scientists and specialized research labs, deep learning is appearing widely in real products. That transition has led to several exciting new deep-learning frameworks that tend to emphasize either rapid prototyping or efficient deployment at scale. For product developers looking to experiment with an idea and then reinforce and deploy it, a single framework that supports both ends of the process is helpful.

Apache MXNet, an open-source deep learning framework first published in 2015, aims to achieve exactly that. I recently talked with Mu Li, principal scientist at Amazon and one of the original authors of MXNet—he was lead author of the “Parameter Server" paper that enables MXNet to have almost linear scale with additional processing power. He walked me through the origins of the framework and his vision of “hybridized” neural network models that can carry a single project from early prototyping through deployment at scale.

MXNet emerged from two goals that are classically opposed in computer science: ease of use and high performance. Declarative structures enable high performance in deep learning: the programmer specifies a network structure at a high level, and the framework implements low-level native routines to build the network in the most efficient manner possible.

The drawback is that these declarative frameworks require the programmer to know and specify the network structure—its computation graph—at the outset. That makes iteration slow and experimentation difficult; discovering the best network structure for a particular problem is arguably the principal task for a deep-learning engineer. In some cases, like long short-term memory (LSTM) networks, the structure of the neural network can depend on control statements--loops and ‘if’ statements--that can’t be evaluated until data is fed in, so programmers need to use equivalent statements provided by the frameworks—meaning lots of mucking about with low-level code.

Declarative frameworks can also be unintuitive to programmers who are accustomed to an interactive trial-and-error approach—an essential shortcoming as interest in deep learning explodes. Many newcomers to deep learning don’t have systematic training in symbolic linear algebra and want to teach themselves through active experimentation.

The deep learning community has responded by introducing a handful of imperative frameworks—notably PyTorch and Chainer—that execute programs line-by-line and allow programmers to use complex control statements to change network structure programmatically and on-the-fly. The drawback here is in performance: if the programmer doesn’t specify the full structure of the network before running it, the framework can’t pre-compile it to wring the best possible performance out of specialized hardware accelerators like GPUs.

“The problem [with imperative programming] is that it’s hard to optimize, because you never know what you’re going to write in the next sentence,” says Li. “You don’t know whether results in memory will be re-used, so it’s hard to get performance right. But it’s ideal for people who want to hack some code together for fast prototyping.” The result is that deep learning implementation sometimes gets split into a research stage using imperative frameworks, and a product stage using declarative frameworks. Li points to Facebook as an example: the company supports both PyTorch and Caffe2, a declarative framework, and uses the former for exploration and the latter for products.

Li and the MXNet developers have taken a hybrid approach that supports experimenting and prototyping with imperative programs, then seamlessly refactoring critical sections into declarative blocks as you move toward production at scale. “We want to have a single interface for users,” says Li. “You can start with imperative code, and then when you want to deploy, you hybridize.”

Developers will have more opportunities to run fast neural networks on all sorts of devices in the next several years. Researchers today depend on computers or cloud services with expensive GPUs to accelerate neural network training. NVIDIA’s GPUs are dominant in this field not just due to raw speed but because CUDA, its programming interface for GPUs, and cuDNN, its deep learning library, are broadly supported by deep learning frameworks, which compile neural networks into code that runs efficiently on NVIDIA GPUs by calling functions provided by cuDNN.

Researchers envision a future in which neural networks will conduct both inference and training on “edge” devices like mobile phones and embedded systems. With this variety of devices in play, from high-end phones to stripped-down IoT devices (and let’s not forget legacy equipment—many of today’s devices will still be in use for much of the coming decade), these networks won’t be able to depend on high-end GPUs; instead, they may use purpose-designed field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs)—or even rely on currently available digital signal processors (DSPs).

That means a flowering of different devices that deep-learning frameworks need to support. The framework authors can’t write extensions for every conceivable accelerator; instead, they’re focusing on general-purpose compilers that can recognize any accelerator hardware and compile neural networks to run efficiently on it.

In the next month or so, MXNet will launch its general compiler, allowing developers to implement neural networks on any accelerator, from high-end GPUs down to inexpensive DSPs and specialized processors in mobile phones. Li says the compiler will initially be open-sourced for early users.

This post is part of a collaboration between O'Reilly and Amazon. See our statement of editorial independence.

Continue reading From prototype to product with hybrid neural networks.

http://ift.tt/2vWYhkn

Four short links: 31 July 2017

Statistics & Fiction, Staying Anonymous, Attacking Machine Learning, and Digital Native is Fiction

The Heretical Things Statistics Tell Us About Fiction (New Yorker) -- Almost without fail, the words evoke their authors’ affinities and manias. John Cheever favors “venereal”—a perfect encapsulation of his urbane midcentury erotics, tinged with morality. Isaac Asimov prefers “terminus,” a word ensconced in a swooping, stately futurism; Woolf has her “mantelpiece,” Wharton her “compunction.” (Melville’s “sperm” is somewhat misleading, perhaps, when separated from his whales.)
Things Not to Do When Anonymous -- so many ways to surrender anonymity, so difficult to remain anonymous.
Robust Physical-World Attacks on Machine Learning Models -- Our algorithm can create spatially constrained perturbations that mimic vandalism or art to reduce the likelihood of detection by a casual observer. We show that adversarial examples generated by RP2 achieve high success rates under various conditions for real road sign recognition by using an evaluation methodology that captures physical world conditions. We physically realized and evaluated two attacks, one that causes a stop sign to be misclassified as a speed limit sign in 100% of the testing conditions, and one that causes a right turn sign to be misclassified as either a stop or added lane sign in 100% of the testing conditions.
The Digital Native is a Myth (Nature) -- The younger generation uses technology in the same ways as older people—and is no better at multitasking.

Continue reading Four short links: 31 July 2017.

http://ift.tt/2tWgEbG

HomePod firmware references new iPhone screen design that has notch cut out at the top, presumably for sensors; a face detection feature and an infrared sensor (Eric Slivka/MacRumors)

Eric Slivka / MacRumors:
HomePod firmware references new iPhone screen design that has notch cut out at the top, presumably for sensors; a face detection feature and an infrared sensor — Late last week, Apple released early firmware for its HomePod smart speaker, which won't be launching to the public until December.

http://ift.tt/2vkQ8c8

German court rules that use of keylogger software to monitor employees is illegal, unless there is concrete suspicion of a criminal offense or breach of duties (The Local)

The Local:
German court rules that use of keylogger software to monitor employees is illegal, unless there is concrete suspicion of a criminal offense or breach of duties — DPA/The Local — Are bosses going too far when they use spy software to track employees' every keystroke?

http://ift.tt/2f00WFN

Sources: Some directors worry Travis Kalanick is trying to game the outcome in favor of his return after he told several people he was "Steve Jobs-ing it" (Kara Swisher/Recode)

Kara Swisher / Recode:
Sources: Some directors worry Travis Kalanick is trying to game the outcome in favor of his return after he told several people he was “Steve Jobs-ing it” — As the car-hailing company searches for cohesion, the former CEO has told some he is “Steve Jobs-ing it.”

http://ift.tt/2hflCdD

eMarketer: Roku will have 38.9M MAUs in 2017, up 19% YoY, compared to 36.9M for Chromecast, 35.8M for Amazon Fire TV, and 21.3M for Apple TV (Sarah Perez/TechCrunch)

Sarah Perez / TechCrunch:
eMarketer: Roku will have 38.9M MAUs in 2017, up 19% YoY, compared to 36.9M for Chromecast, 35.8M for Amazon Fire TV, and 21.3M for Apple TV — Roku may be the most-used connected TV device this year, edging out Chromecast, Fire TV and Apple TV, according to a new 2017 forecast from eMarketerreleased this morning.

http://ift.tt/2tVDT5M

Sunday, July 30, 2017

These New Devices Promise to Fight Pain Without Opioids

These New Devices Promise to Fight Pain Without Opioids

Their goal is to replace addictive painkillers and help people detox from opioids.

http://ift.tt/2heddHw

Behind the chart-topping anonymous messaging app Sarahah, which gives message recipients no way to respond (Tim Bradshaw/Financial Times)

Tim Bradshaw / Financial Times:
Behind the chart-topping anonymous messaging app Sarahah, which gives message recipients no way to respond — Text service tops iPhone download charts, but the real challenge is staying there — Read next … There are two unusual things about Sarahah, a mobile app topping iPhone App Store download charts around the world.

http://ift.tt/2ubedND

Samsung DeX - Geek News Central

The Samsung DeX Station converts a Galaxy S8 or S8+ smartphone into a desktop computer. Plop the S8 in the DeX, plug in a keyboard and mouse, hook up a TV, and you’re set with Android on the big screen. That’s the theory, what about in practice? Let’s take a look.

The DeX hardware is circular device, about 10 cm across, with a top surface that sweeps smoothly back and up to reveal the USB C connector for the smartphone.

Around the base are two USB 2 ports, a full-size HDMI socket, a 100 Mb/s network connector and a USB C for powering the DeX.

Getting setup is simplicity itself. Connect all the hardware up and slot the phone in. There’s no additional software to add as it’s all built-in to the S8 and the DeX itself. I used a wireless keyboard and mouse combo connecting to a USB transmitter. The TV connected to the HDMI port with a cable and I used the WiFi on the S8 for networking.

When the S8 is placed in the DeX, a prompt appears asking whether to start DeX or to only mirror the S8 screen. Choosing the former gets the DeX desktop in its full HD glory and it looks convincing. But what’s it really like?

Let’s start with the positives…the DeX desktop is what you’d expect an Android desktop to be like, using familiar apps in a windowed world. It’s fast and all the apps on the phone are available through DeX. Google Maps works and it’s perfect for YouTube and web browsing. Samsung promotes DeX-optimised apps via its app store.

But while many apps seemed to be quite happy with DeX and run in both full screen and windowed modes, some apps don’t like DeX and display as if they are on the S8 in portrait. This is frustrating and while this could be expected for games like Monument Valley, it seems odd that Netflix can’t cope – surely this would be seen as a “must have” by Samsung? Some apps don’t have all the necessary controls either – it’s tricky to pinch-to-zoom with only a mouse pointer.

Of course, games players and movie watchers aren’t the target audience for DeX. Samsung see this as a tool for business and promote the benefits of Microsoft apps and Office365 in the literature. For example, instead of a lugging a laptop for a presentation, take DeX, plug it into the data projector and you’re sorted. Need to do a quick bit of editing? Steal a desk, connect up DeX, fire up Word and you’re working.

Where DeX also scores well is with VDIs (Virtual Desktop Infrastructures) like Citrix. Connect through to your office backend to run a virtual Windows PC and you can be working as if you are at your own desk. From that point of view, it’s slick. While DeX isn’t going to replace a laptop on an extended business trip, it makes sense for a short visit when you want to travel light.

Pricewise, the DeX station has an RRP of GB£129, though it’s already discounted by £40 in several online stores. It’s still a percentage on top of the S8 and S8+.

Overall, DeX does what it sets out to do and the key question here is not about the technology. It’s whether Samsung’s vision and the DeX Station match your way of working. YMMV, as they say.

Samsung DeX

http://ift.tt/2vkkag9

Monday, July 31, 2017

Sunday, July 30, 2017

Check us out today!