Google is baking machine learning into its BigQuery data warehouse

There are still a lot of obstacles to building machine learning models and one of those is that in order to build those models, developers often have to move a lot of data back and forth between their data warehouses and wherever they are building their models. Google is now making this part of the process a bit easier for the developers and data scientists in its ecosystem with BigQuery ML, a new feature of its BigQuey data warehouse by building some machine learning functionality right into BigQuery.

Using BigQuery ML, developers can build models using linear and logistical regression right inside their data warehouse without having to transfer data back and forth as they build and fine-tune their models. And all they have to do to build these models and get predictions is to write a bit of SQL.

Moving data doesn’t sound like it should be a big issue, but developers often spend a lot of their time on this kind of grunt work — time that would be better spend on actually working on their models.

BigQuery ML also promises to make it easier to build these models, even for developers who don’t have a lot of experience with machine learning. To get started, developers can use what’s basically a variant of standard SQL to say what kind of model they are trying to build and what the input data is supposed to be. From there, BigQuery ML then builds the model and allows developers to almost immediately generate predictions based on it. And they won’t even have to write any code in R or Python.

These new features are now available in beta.

Read Full Article

Google is making a fast specialized TPU chip for edge devices and a suite of services to support it

In a pretty substantial move into trying to own the entire AI stack, Google today announced that it will be rolling out a version of its Tensor Processing Unit — a custom chip optimized for its machine learning framework TensorFlow — optimized for inference in edge devices.

That’s a bit of a word salad to unpack, but here’s the end result: Google is looking to have a complete suite of customized hardware for developers looking to build products around machine learning, such as image or speech recognition, that it owns from the device all the way through to the server. Google will have the cloud TPU (the third version of which will soon roll out) to handle training models for various machine learning-driven tasks, and then run the inference from that model on a specialized chip that runs a lighter version of TensorFlow that doesn’t consume as much power. Google is exploiting an opportunity to split the process of inference and machine training into two different sets of hardware and dramatically reduce the footprint required in a device that’s actually capturing the data. That would result in faster processing, less power consumption, and potentially more importantly, a dramatically smaller surface area for the actual chip.

Google is also rolling out a new set of services to compile TensorFlow (Google’s machine learning development framework) into a lighter-weight version that can run on edge devices without having to call the server for those operations. That, again, reduces the latency and could have any number of results, from safety (in autonomous vehicles) to just a better user experience (voice recognition). As competition heats up in the chip space, both from the larger companies and from the emerging class of startups, nailing these use cases is going to be really important for larger companies. That’s especially true for Google as well, which also wants to own the actual development framework in a world where there are multiple options like Caffe2 and PyTorch.

Google will be releasing the chip on a kind of modular board not so dissimilar to the Raspberry Pi, which will get it into the hands of developers that can tinker and build unique use cases. But more importantly, it’ll help entice developers who are already working with TensorFlow as their primary machine learning framework with the idea of a chip that’ll run those models even faster and more efficiently. That could open the door to new use cases and ideas, and should it be successful, will lock those developers further into Google’s cloud ecosystem on both the hardware (the TPU) and framework (TensorFlow) level. While Amazon owns most of the stack for cloud computing (with Azure being the other largest player), it looks like Google is looking to own the whole AI stack – and not just offer on-demand GPUs as a stopgap to keep developers operating within that ecosystem.

Thanks to the proliferation of GPUs, machine learning has become increasingly common across a variety of use cases, which doesn’t just require the horsepower to train a model to identify what a cat looks like. It also needs the ability to take in an image and quickly identify that said four-legged animal is a cat based on the model it’s trained with tens of thousands (or more) images of what a cat is. GPUs were great for both use cases, but it’s clear that better hardware is necessary with the emergence of use cases like autonomous driving, photo recognition on cameras, or a variety of others — for which even millisecond-level lag is too much and power consumption, or surface area, is a dramatic limiting factor.

The edge-specialized TPU is an ASIC chip, a breed of chip architecture that’s increasingly popular for specific use cases like mining for cryptocurrency (such as larger companies like Bitmain). The chips excel at doing specific things really well, and it’s opened up an opportunity to tap various niches, such as mining cryptocurrency, with specific chips that are optimized for those calculations. These kinds of edge-focused chips tend to do a lot of low-precision calculations very fast, making the whole process of juggling runs between memory and the actual core significantly less complicated and consuming less power as a result.

While Google’s entry into this arena has long been a whisper in the Valley, this is a stake in the ground for the company that it wants to own everything from the hardware all the way up to the end user experience, passing through the development layer and others on the way there. It might not necessarily alter the calculus of the ecosystem, as even though it’s on a development board to create a playground for developers, Google still has to make an effort to get the hardware designed into other pieces of hardware and not just its own if it wants to rule the ecosystem. That’s easier said than done, even for a juggernaut like Google, but it is a big salvo from the company that could have rather significant ramifications down the line as every big company races to create its own custom hardware stack that’s specialized for its own needs.

Read Full Article

Google brings its search technology to the enterprise

One of Google’s first hardware products was its search appliance, a custom-built server that allowed businesses to bring Google’s search tools to the data behind their firewalls. That appliance is no more, but Google today announced the spiritual successor to it with an update to Cloud Search. Until today, Cloud Search only indexed G Suite data. Now, it can pull in data from a wide variety of third-party services that can run on-premise or in the cloud, too, making the tool far more useful for large businesses that want to make all of their data searchable by their employees.

“We are essentially taking all of Google expertise in search and are applying it to your enterprise content,” Google said.

One of the launch customers for this new service is Whirlpool, which built its own search portal and indexed over 12 million documents from more than a dozen services using this new service.

“This is about giving employees access to all the information from across the enterprise, even if it’s traditionally siloed data whether that’s in a database or a legacy productivity tool and make all of that available in a single index,” Google explained.

To enable this functionality, Google is making a number of software adapters available that will bridge the gap between these third-party services and Cloud Search. Over time, Google wants to add support for more services and bring this cloud-based technology on par with what its search appliance was once capable of.

Read Full Article

Google Cloud introduces shielded VMs for additional security

While we might like to think all of our applications are equal in our eyes, in reality some are more important than others and require an additional level of security. To meet those requirements, Google introduced shielded virtual machines at Google Next today.

As Google describes it, “Shielded VMs leverage advanced platform security capabilities to help ensure your VMs have not been tampered with. With Shielded VMs, you can monitor and react to any changes in the VM baseline as well as its current runtime state.”

These specialized VMs run on GCP and come with a set of partner security controls to defend against things like rootkits and bootkits, according to Google. There are a whole bunch of things that happen even before an application launches inside a VM, and each step in that process is vulnerable to attack.

That’s because as the machine starts up, before you even get to your security application, it launches the firmware, the boot sequence, the kernel, then the operating system — and then and only then, does your security application launch.

That time between startup and the security application launching could leave you vulnerable to certain exploits that take advantage of those openings. The shielded VMs strip out as much of that process as possible to reduce the risk.

“What we’re doing here is we are stripping out any of the binary that doesn’t absolutely have to be there. We’re ensuring that every binary that is there is signed, that it’s signed by the right party, and that they load in the proper sequence,” a Google spokesperson explained. All of these steps should reduce overall risk.

Shielded VMs are available in Beta now

Read Full Article

Google introduces ‘Context-aware’ access to supplement traditional logons

We know by now that usernames and passwords are a poor way of securing applications and online services, but they remain for the most part a key tool in the security arsenal. The trouble is that with all of the security breaches in recent years from Equifax to Anthem to Target (and many others), people’s credentials have been widely shared on the internet black market.

Google wants to help fix that problem and today at Google Next, it announced Context-aware access, a new program that looks beyond your credentials to other factors to help determine if it’s really you or someone pretending to be you.

Context-aware access lets administrators define a set of information that could help them more accurately ascertain the identity of the person trying to access your service. “Context-aware access allows organizations to define and enforce granular access to GCP APIs, resources, G Suite, and third-party SaaS apps based on a user’s identity, location, and the context of their request,” Google explained.

One way to better understand the person accessing your services is to look at some contextual clues such as where they are logging on, the IP address of the machine they are logging on from, the time of day and other factors. Does all of this make sense based on what you know about the person?

The idea flips the notion of security responsibility on its head. Instead of requiring the user to be completely responsible for proving who they are, it puts the burden (and control) in the hands of the administrator where it makes more sense.

Google created this security tool because it recognizes as well as any company that users aren’t tied to the office anymore. They are working on mobile devices and accessing apps and cloud services and it becomes more difficult to trust identity, especially with so many stolen credentials out there.

The new program builds on Google’s BeyondCorp vision, an idea they began developing in 2011 to address the fact that computing no longer took place inside a clearly defined perimeter. In the days before mobile and cloud, people generally accessed computer systems from a specific place. If someone tried to log on from outside of that, you could catch them and turn them away.

Mobile and the cloud changed all that and Google began defining an idea called Zero Trust, the notion that you don’t trust anyone on your services and build an appropriate security position based on that idea. Identity is really central to this, but at some point, even in a Zero Trust model, you have to let people in and do business on your services. This tool gives administrators in a Zero Trust Model much more information to work with beyond a username and password to determine if the user is authentic or not.

Context-aware access management is available today for customers using ​VPC Service Controls. It will be available soon for customers using ​Cloud Identity and Access Management​ (IAM), ​Cloud Identity-Aware Proxy​ (IAP), and ​Cloud Identity​, according to the company.

Read Full Article

How to Turn a TV Into a Retro Gaming System With Raspberry Pi Zero

Facebook’s chief legal officer to leave this year

Facebook’s chief legal officer Colin Stretch has announced he’ll be out by the end of the year. 

In the inevitable Facebook post explaining why he’s moving on, Stretch writes that after he and his wife made a decision to move back to DC from California “a few years ago… we knew it would be difficult for me to remain in this role indefinitely”.

“As Facebook embraces the broader responsibility Mark [Zuckerberg] has discussed in recent months, I’ve concluded that the company and the Legal team need sustained leadership in Menlo Park,” he adds, saying he’ll stay to the end of the year to help with the transition.

Facebook has had a very awkward two years so far as politically charged scandals go. First revelations about the massive Kremlin-fueled election interference which it totally missed. Then the massive Cambridge Analytica data misuse debacle which Facebook also claims to have totally missed, even though it (still apparently) employs one of the academics whose quiz app was the vehicle used to suck out people’s data.

Since then a bunch of follow-on admissions have flowed from the company confirming that access to user data on its platform wasn’t as locked down as it’s historically liked to claim — albeit, despite masses of evidence to the contrary.

Nor, perhaps, as the FTC might have expected give a 2011 privacy settlement with the company. The regulator has now opened a fresh investigation. Meanwhile Facebook is carrying out a retrospective app audit — a not so tacit admission about its abject lack of enforcement of its own developer policy.

And yet there have not — at least publicly — been any heads rolling at Facebook despite all this failure.

Most likely because, as founder Mark Zuckerberg recently told Recode’s Kara Swisher during a podcast interview: “I designed the platform, so if someone’s going to get fired for this, it should be me.”

Of course Zuckerberg isn’t going to fire himself. Not when he doesn’t have to. Given the structure of the company he’s sitting pretty on his CEO throne, no matter how tarnished that crown now is.

Instead of firing himself — let’s not forget his 2016 attempts to dismiss the notion of Facebook-enabled election interference as a “pretty crazy idea” — Zuckerberg once again fired up his multi-year apology tour for privacy and data-related screw ups, rolling this through 2017 and  2018, as fresh scandals rocked the company’s reputation. And raised the specter of regulation to control damaging activity on the platform that the company has spectacularly failed to control.

Though you’d be hard pressed to read any of this scandalabra just by looking at the company’s earnings and stock price. Perhaps because investors view any regulation as likely to cement Facebook’s dominance, rather than upset the apple cart in a way that could allow a younger model to come in and disrupt its grip on consumers’ eyeballs.

Even so, 2018 has seen Zuckerberg, if not literally dragged but politically compelled to appear in front of US and EU lawmakers — where he faced a barrage of questions; some dumb, others cutting to the heart of the company’s contradictions and its contradictory claims.

Last year Facebook’s chief legal officer Colin Stretch was also in the Senate, alongside reps from Google and Twitter, fielding awkward questions about Russian election interference and the spread of extremist content on the platform.

There Stretch made an unfortunate slip of the tongue during his introductory remarks — seemingly saying “keeping people unsafe on Facebook is critical to our mission” before quickly correcting himself to stress he’d meant to say “keeping people safe”. As Freudian slips go it’s a doozy.

But it’s certainly not a great time for Facebook to be losing its general counsel. Not with so much ongoing political and legal risk. Although if Zuckerberg isn’t going to go then perhaps other Facebook veterans will feel compelled to leave on his behalf.

With the usual departing platitudes, Stretch writes: “This has not been an easy decision. Companies are made up of people, and the people here are talented, caring, and most of all committed to doing the right thing. Even now, eight-and-a-half years after I started, I often stop myself and ask how I got so lucky to be a part of this.”

“There is never a ‘right time’ for a transition like this, but the team and the company boast incredible talent and will navigate this well,” he adds.

In March it also emerged that Facebook would likely be parting ways with its long-time chief security officer, Alex Stamos, this summer — after the New York Times reported on internal disagreements between the CSO and other  execs, saying Stamos had wanted Facebook to be more public about the misuse of its platform by nation states.

This week BuzzFeed News obtained an internal memo sent by Stamos in March, days after he had confirmed his plans to leave the company, in which he writes: “I was the Chief Security Officer during the 2016 election season, and I deserve as much blame (or more) as any other exec at the company.”

Though he demurs on confirming whether he has actually quit for real at that point — but does admit to having had “passionate discussions with other execs”, including, seemingly, about Facebook’s approach to sharing public data on Russian disinformation.

“The world has changed from underneath us in many ways. One change has been the thrusting of private tech companies into the struggle between nation-states,” he writes on this. “Traditionally, the standard has been to report malicious activity by adversary nations to US law enforcement. We are moving into a world where the major platforms are going to be expected to provide our findings, attribution and data directly to the public, making us a visible participant in the battle between cyberwarfare titans.”

“This is an uncomfortable transition, and have not always agree with the compromises we have struck in the process. That being said, I believe my colleagues have all approached the process in good faith, and together we have sorted through legitimate equities that needed to be weighed,” Stamos adds.

Stamos goes on to implore colleagues to make major changes “to win back the world’s trust” — including rethinking the metrics Facebook fixes itself to as a business; being more adversarial in its thinking when building products and processes; and — in what looks very much like a swipe at the company’s use of dark pattern design in its consent flows — re-engineering how it gathers user data to be more honest and minimize (rather than maximize) data collection.

On that it’s worth noting that privacy by design is a core plank of Europe’s new data protection framework, GDPR — which Stamos is seemingly describing at one point in the memo, without giving it a literal name-check.

“We need to build a user experience that conveys honesty and respect, not one optimized to get people to click yes to giving us more access. We need to intentionally not collect data where possible, and to keep it only as long as we are using it to serve people,” he writes [emphasis his]. “We need to find and stop adversaries who will be copying the playbook they saw in 2016. We need to listen to people (including internally) when they tell us a feature is creepy or point out a negative impact we are having in the world. We need to deprioritze short-term growth and revenue and to explain to Wall Street why that is ok. We need to be willing to pick sides when there are clear moral or humanitarian issues. And we need to be open, honest and transparent about challenges and what we are doing to fix them.”

Read Full Article

The 10 Best iPhone X Cases to Protect, Charge, and Show Off

Grubhub acquires payments and loyalty company LevelUp for $390M

Grubhub announced this morning that it’s agreed to acquire LevelUp for $390 million cash.

Founder and CEO Matt Maloney told me that while previous Grubhub acquisitions like Eat24 were designed to give the company’s delivery business more scale, “This is kind of a different acquisition. It’s a product and strategic positioning acquisition.”

LevelUp is based in Boston and offers a platform to manage digital ordering, payments and loyalty. Maloney said that buying the company allows Grubhub to deepen its integration with restaurants’ point-of-sale systems. That, in turn, will allow them to handle more deliveries.

At the same time, Maloney said LevelUp can help Grubhub build a restaurant platform that goes beyond delivery, for example by managing their customer interactions across mobile and the web.

“We want to help restaurants actively engage with their diners,” Maloney said. “This is a huge step in that direction.”

Once the regulatory waiting period is over, the entire team LevelUp team will be joining Grubhub, with founder and CEO Seth Priebatsch reporting to Maloney — who said that in the short term, he plans to change very little, aside from the POS integrations. Even in the long term, he suggested that LevelUp could continue to operate as its own brand within the larger Grubhub platform.

“They’re doing something really well and we don’t want to screw that up,” he said. “We want to make as little change as possible, until we all understand how we’re better working together.”

The LevelUp platform was launched in 2011, and the company has raised around $108 million in total funding, according to Crunchbase. Investors include Highland Capital, GV, Balderton Capital, Deutsche Telecom Strategic Investments, Continental Advisors, Transmedia Capital and U.S. Boston Capital.

“For the last seven years, we have worked to provide restaurant clients with a complete solution to engage customers, and this agreement is the biggest and most exciting step in achieving that mission,” Priebatsch said in a statement provided by Grubhub. “After close, the entire team will remain in Boston and our office will become Grubhub’s newest center of technology excellence.”

The announcement came as part of Grubhub’s second quarter earnings release, which saw the company grow active diners by 70 percent year-over-year, to 15.6 million, while revenue increased 51 percent, to $240 million.

Read Full Article

Will Upgrading Your Old Computer Make It Use Less Power?

SpaceX lands Falcon 9 booster on Just Read The Instructions drone ship

SpaceX confirmed on Twitter this morning that it recovered the booster from the latest Falcon 9 launch. Shortly after launching from Vandenberg Air Force Base in Southern California at 7:39AM ET this morning, the booster stage landed on the Just Read The Instructions drone ship. The company will now try to catch the rocket’s fairing with a giant net attached to the ship Mr. Stevens.

Despite challenging weather conditions, Falcon 9 first stage booster landed on Just Read the Instructions.

— SpaceX (@SpaceX) July 25, 2018

SpaceX has become more adept at landing its booster rockets but it’s still a spectacle every time it happens. This landing is extra special as the winds were gusting around the time of the launch.

The rocket company has so far been less successful with catching the payload shrouds. SpaceX’s high-speed recovery boat Mr. Steven took to the seas this time around with a larger net in the hopes of recovering the fairings. Reusing as much as possible is critical to SpaceX’s mission to lower the cost of space flight.

 

Developing…

Read Full Article

Computer vision researchers build an AI benchmark app for Android phones

A group of computer vision researchers from ETH Zurich want to do their bit to enhance AI development on smartphones. To wit: They’ve created a benchmark system for assessing the performance of several major neural network architectures used for common AI tasks.

They’re hoping it will be useful to other AI researchers but also to chipmakers (by helping them get competitive insights); Android developers (to see how fast their AI models will run on different devices); and, well, to phone nerds — such as by showing whether or not a particular device contains the necessary drivers for AI accelerators. (And, therefore, whether or not they should believe a company’s marketing messages.)

The app, called AI Benchmark, is available for download on Google Play and can run on any device with Android 4.1 or higher — generating a score the researchers describe as a “final verdict” of the device’s AI performance.

AI tasks being assessed by their benchmark system include image classification, face recognition, image deblurring, image super-resolution, photo enhancement or segmentation.

They are even testing some algorithms used in autonomous driving systems, though there’s not really any practical purpose for doing that at this point. Not yet anyway. (Looking down the road, the researchers say it’s not clear what hardware platform will be used for autonomous driving — and they suggest it’s “quite possible” mobile processors will, in future, become fast enough to be used for this task. So they’re at least prepped for that possibility.)

The app also includes visualizations of the algorithms’ output to help users assess the results and get a feel for the current state-of-the-art in various AI fields.

The researchers hope their score will become a universally accepted metric — similar to DxOMark that is used for evaluating camera performance — and all algorithms included in the benchmark are open source. The current ranking of different smartphones and mobile processors is available on the project’s webpage.

The benchmark system and app was around three months in development, says AI researcher and developer Andrey Ignatov.

He explains that the score being displayed reflects two main aspects: The SoC’s speed and available RAM.

“Let’s consider two devices: one with a score of 6000 and one with a score of 200. If some AI algorithm will run on the first device for 5 seconds, then this means that on the second device this will take about 30 times longer, i.e. almost 2.5 minutes. And if we are thinking about applications like face recognition this is not just about the speed, but about the applicability of the approach: Nobody will wait 10 seconds till their phone will be trying to recognize them.

“The same is about memory: The larger is the network/input image — the more RAM is needed to process it. If the phone has small amount of RAM that is e.g. only enough to enhance 0.3MP photo, then this enhancement will be clearly useless, but if it can do the same job for Full HD images — this opens up much wider possibilities. So, basically the higher score — the more complex algorithms can be used / larger images can be processed / it will take less time to do this.”

Discussing the idea for the benchmark, Ignatov says the lab is “tightly bound” to both research and industry — so “at some point we became curious about what are the limitations of running the recent AI algorithms on smartphones”.

“Since there was no information about this (currently, all AI algorithms are running remotely on the servers, not on your device, except for some built-in apps integrated in phone’s firmware), we decided to develop our own tool that will clearly show the performance and capabilities of each device,” he adds. 

“We can say that we are quite satisfied with the obtained results — despite all current problems, the industry is clearly moving towards using AI on smartphones, and we also hope that our efforts will help to accelerate this movement and give some useful information for other members participating in this development.”

After building the benchmarking system and collating scores on a bunch of Android devices, Ignatov sums up the current situation of AI on smartphones as “both interesting and absurd”.

For example, the team found that devices running Qualcomm chips weren’t the clear winners they’d imagined — i.e. based on the company’s promotional materials about Snapdragon’s 845 AI capabilities and 8x performance acceleration.

“It turned out that this acceleration is available only for ‘quantized’ networks that currently cannot be deployed on the phones, thus for ‘normal’ networks you won’t get any acceleration at all,” he says. “The saddest thing is that actually they can theoretically provide acceleration for the latter networks too, but they just haven’t implemented the appropriated drivers yet, and the only possible way to get this acceleration now is to use Snapdragon’s proprietary SDK available for their own processors only. As a result — if you are developing an app that is using AI, you won’t get any acceleration on Snapdragon’s SoCs, unless you are developing it for their processors only.”

Whereas the researchers found that Huawei’s Kirin’s 970 CPU — which is technically even slower than Snapdragon 636 — offered a surprisingly strong performance.

“Their integrated NPU gives almost 10x acceleration for Neural Networks, and thus even the most powerful phone CPUs and GPUs can’t compete with it,” says Ignatov. “Additionally, Huawei P20/P20 Pro are the only smartphones on the market running Android 8.1 that are currently providing AI acceleration, all other phones will get this support only in Android 9 or later.”

It’s not all great news for Huawei phone owners, though, as Ignatov says the NPU doesn’t provide acceleration for ‘quantized’ networks (though he notes the company has promised to add this support by the end of this year); and also it uses its own RAM — which is “quite limited” in size, and therefore you “can’t process large images with it”…

“We would say that if they solve these two issues — most likely nobody will be able to compete with them within the following year(s),” he suggests, though he also emphasizes that this assessment only refers to the one SoC, noting that Huawei’s processors don’t have the NPU module.

For Samsung processors, the researchers flag up that all the company’s devices are still running Android 8.0 but AI acceleration is only available starting from Android 8.1 and above. Natch.

They also found CPU performance could “vary quite significantly” — up to 50% on the same Samsung device — because of throttling and power optimization logic. Which would then have a knock on impact on AI performance.

For Mediatek, the researchers found the chipmaker is providing acceleration for both ‘quantized’ and ‘normal’ networks — which means it can reach the performance of “top CPUs”.

But, on the flip side, Ignatov calls out the company’s slogan — that it’s “Leading the Edge-AI Technology Revolution” — dubbing it “nothing more than their dream”, and adding: “Even the aforementioned Samsung’s latest Exynos CPU can slightly outperform it without using any acceleration at all, not to mention Huawei with its Kirin’s 970 NPU.”

“In summary: Snapdragon — can theoretically provide good results, but are lacking the drivers; Huawei — quite outstanding results now and most probably in the nearest future; Samsung — no acceleration support now (most likely this will change soon since they are now developing their own AI Chip), but powerful CPUs; Mediatek — good results for mid-range devices, but definitely no breakthrough.”

It’s also worth noting that some of the results were obtained on prototype samples, rather than shipped smartphones, so haven’t yet been included in the benchmark table on the team’s website.

“We will wait till the devices with final firmware will come to the market since some changes might still be introduced,” he adds.

Read Full Article

Computer vision researchers build an AI benchmark app for Android phones

A group of computer vision researchers from ETH Zurich want to do their bit to enhance AI development on smartphones. To wit: They’ve created a benchmark system for assessing the performance of several major neural network architectures used for common AI tasks.

They’re hoping it will be useful to other AI researchers but also to chipmakers (by helping them get competitive insights); Android developers (to see how fast their AI models will run on different devices); and, well, to phone nerds — such as by showing whether or not a particular device contains the necessary drivers for AI accelerators. (And, therefore, whether or not they should believe a company’s marketing messages.)

The app, called AI Benchmark, is available for download on Google Play and can run on any device with Android 4.1 or higher — generating a score the researchers describe as a “final verdict” of the device’s AI performance.

AI tasks being assessed by their benchmark system include image classification, face recognition, image deblurring, image super-resolution, photo enhancement or segmentation.

They are even testing some algorithms used in autonomous driving systems, though there’s not really any practical purpose for doing that at this point. Not yet anyway. (Looking down the road, the researchers say it’s not clear what hardware platform will be used for autonomous driving — and they suggest it’s “quite possible” mobile processors will, in future, become fast enough to be used for this task. So they’re at least prepped for that possibility.)

The app also includes visualizations of the algorithms’ output to help users assess the results and get a feel for the current state-of-the-art in various AI fields.

The researchers hope their score will become a universally accepted metric — similar to DxOMark that is used for evaluating camera performance — and all algorithms included in the benchmark are open source. The current ranking of different smartphones and mobile processors is available on the project’s webpage.

The benchmark system and app was around three months in development, says AI researcher and developer Andrey Ignatov.

He explains that the score being displayed reflects two main aspects: The SoC’s speed and available RAM.

“Let’s consider two devices: one with a score of 6000 and one with a score of 200. If some AI algorithm will run on the first device for 5 seconds, then this means that on the second device this will take about 30 times longer, i.e. almost 2.5 minutes. And if we are thinking about applications like face recognition this is not just about the speed, but about the applicability of the approach: Nobody will wait 10 seconds till their phone will be trying to recognize them.

“The same is about memory: The larger is the network/input image — the more RAM is needed to process it. If the phone has small amount of RAM that is e.g. only enough to enhance 0.3MP photo, then this enhancement will be clearly useless, but if it can do the same job for Full HD images — this opens up much wider possibilities. So, basically the higher score — the more complex algorithms can be used / larger images can be processed / it will take less time to do this.”

Discussing the idea for the benchmark, Ignatov says the lab is “tightly bound” to both research and industry — so “at some point we became curious about what are the limitations of running the recent AI algorithms on smartphones”.

“Since there was no information about this (currently, all AI algorithms are running remotely on the servers, not on your device, except for some built-in apps integrated in phone’s firmware), we decided to develop our own tool that will clearly show the performance and capabilities of each device,” he adds. 

“We can say that we are quite satisfied with the obtained results — despite all current problems, the industry is clearly moving towards using AI on smartphones, and we also hope that our efforts will help to accelerate this movement and give some useful information for other members participating in this development.”

After building the benchmarking system and collating scores on a bunch of Android devices, Ignatov sums up the current situation of AI on smartphones as “both interesting and absurd”.

For example, the team found that devices running Qualcomm chips weren’t the clear winners they’d imagined — i.e. based on the company’s promotional materials about Snapdragon’s 845 AI capabilities and 8x performance acceleration.

“It turned out that this acceleration is available only for ‘quantized’ networks that currently cannot be deployed on the phones, thus for ‘normal’ networks you won’t get any acceleration at all,” he says. “The saddest thing is that actually they can theoretically provide acceleration for the latter networks too, but they just haven’t implemented the appropriated drivers yet, and the only possible way to get this acceleration now is to use Snapdragon’s proprietary SDK available for their own processors only. As a result — if you are developing an app that is using AI, you won’t get any acceleration on Snapdragon’s SoCs, unless you are developing it for their processors only.”

Whereas the researchers found that Huawei’s Kirin’s 970 CPU — which is technically even slower than Snapdragon 636 — offered a surprisingly strong performance.

“Their integrated NPU gives almost 10x acceleration for Neural Networks, and thus even the most powerful phone CPUs and GPUs can’t compete with it,” says Ignatov. “Additionally, Huawei P20/P20 Pro are the only smartphones on the market running Android 8.1 that are currently providing AI acceleration, all other phones will get this support only in Android 9 or later.”

It’s not all great news for Huawei phone owners, though, as Ignatov says the NPU doesn’t provide acceleration for ‘quantized’ networks (though he notes the company has promised to add this support by the end of this year); and also it uses its own RAM — which is “quite limited” in size, and therefore you “can’t process large images with it”…

“We would say that if they solve these two issues — most likely nobody will be able to compete with them within the following year(s),” he suggests, though he also emphasizes that this assessment only refers to the one SoC, noting that Huawei’s processors don’t have the NPU module.

For Samsung processors, the researchers flag up that all the company’s devices are still running Android 8.0 but AI acceleration is only available starting from Android 8.1 and above. Natch.

They also found CPU performance could “vary quite significantly” — up to 50% on the same Samsung device — because of throttling and power optimization logic. Which would then have a knock on impact on AI performance.

For Mediatek, the researchers found the chipmaker is providing acceleration for both ‘quantized’ and ‘normal’ networks — which means it can reach the performance of “top CPUs”.

But, on the flip side, Ignatov calls out the company’s slogan — that it’s “Leading the Edge-AI Technology Revolution” — dubbing it “nothing more than their dream”, and adding: “Even the aforementioned Samsung’s latest Exynos CPU can slightly outperform it without using any acceleration at all, not to mention Huawei with its Kirin’s 970 NPU.”

“In summary: Snapdragon — can theoretically provide good results, but are lacking the drivers; Huawei — quite outstanding results now and most probably in the nearest future; Samsung — no acceleration support now (most likely this will change soon since they are now developing their own AI Chip), but powerful CPUs; Mediatek — good results for mid-range devices, but definitely no breakthrough.”

It’s also worth noting that some of the results were obtained on prototype samples, rather than shipped smartphones, so haven’t yet been included in the benchmark table on the team’s website.

“We will wait till the devices with final firmware will come to the market since some changes might still be introduced,” he adds.

Read Full Article

Market research firm GlobalWebIndex takes first VC with $40M Series A

Market research firm GlobalWebIndex, which provides consumer insight data for marketing purposes for customers including Google, Spotify, WPP, IPG and Omnicom Group, has closed a $40 million Series A round. The funding is its first VC raise, almost a decade after the business was founded.

The investment comes from New York-based growth fund Stripes Group, along with a number of other unnamed data, software and consumer technology companies. GWI says it will be used to accelerate product development and for international expansion, including in the U.S. and Asia. 

The company is based in London but has recently opened offices in New York City and Los Angeles, as well as having technology hubs in locations across Europe.

With the new funding it says it’s planning to open more international offices across the Americas and Asia Pacific to support a client base which spans more than 80 countries.  

But why take VC now? “After nine years with no funding and seeing phenomenal growth and expansion, we are still seeing an increasing demand for our data, especially from companies that we haven’t traditionally sold to,” says CEO Tom Smith.

“The new funding will support our product development and hiring efforts so we can establish ourselves as the go to platform for digital consumer insights for the marketing industry.”

The company believes it’s positioned itself on the right side of digital history, having chosen an opt-in, survey-based route for gathering a chunk of its consumer data for market research purposes — putting detailed questions to its global panel of 22 million web users from whom it’s gaining up front consent to their data being processed.

Europe’s new data protection framework, GDPR, is explicit on the need for consent to be informed, specific and freely given if that’s your legal basis for processing people’s personal data.

On the product development front, GWI says it’s working to develop new ways of collecting consumer data — having developed a proprietary, device-agnostic “messenger-style survey tool” which Smith says allows respondents to “answer questions at times and in formats which suit them”.

“It’s about putting the consumer first — not just in how their data is used, but how you run the survey itself. With this new approach we hope to be able to return survey results faster to our clients, so they can make quick business decisions based on insights retrieved from our tried and tested methodology,” he adds.

“Years before GDPR, we wanted to be respectful to those who take one of our surveys, meaning that they see consumer-centric privacy and consent notices which use easy-to-understand language to outline what we do, why it’s important to our clients, and what their responses will be used for.”

“We ask our respondents a wide range of questions relating to their digital lives and lifestyles,” he continues. “This covers everything from their social media, device, media consumption and online behaviors to their interactions with brands, their attitudes and their daily life. For us, it’s key to understand not just what people are doing online, but the attitudes, motivations and beliefs which drive this.”

GWI’s primary product offering is its core study — which is fielded in 44 countries, and which it says contains 25,000 data points on 22M+ “connected consumers”.

“Each year we interview hundreds of thousands of representative individuals about their digital lives. The aggregated results of this are made available on our industry-leading platform on a syndicated basis, where clients can build and then analyze any audience they like. For example, you could look at Older vs Younger Millennials, Mums vs Dads, Instagrammers vs Snapchatters — in each case, understanding which behaviors and attitudes are unique to each group,” explains Smith.

It also offers a range of custom services to supplement that core survey-based market intelligence data.

“Many clients use this offering to ask follow-up questions to our respondents, allowing them to overlay the answers to their bespoke questions with the 25,000 data points contained in the core data set,” he says, adding: “We also offer a wide range of other research solutions such as brand tracking, path to purchase journey, ad-effectiveness, concept testing, website analytics and more. Here the sheer scale of our panel — currently at 22M consumers — is a real differentiator.”

All survey respondents are compensated for their time, according to Smith — in different ways, depending on the market, but including via monetary payments, vouchers, loyalty points, charitable donations etc.

One thing to note is that GWI does also use cookie-based tracking to gather data less directly — including by working with partners. So it is also reliant on third parties obtaining consent to data processing, and must therefore rely on those partners to cleanly and fairly obtain consent for this portion of its market intelligence activity.

“In some instances, the cookies we use as part of our research are dropped by the research panels we work with, rather than by GlobalWebIndex itself. In these instances, all such panels are required to obtain consent in GDPR-compliant manners,” says Smith on this.

The company tells TechCrunch it uses cookie tracking to enrich its core survey data, and only uses cookies to track the behaviors of its panelists — and only those who have actively opted in to this type of tracking.

“Our analytics technology connects the data we collect through surveys to the behaviors of our panel on client websites and connected properties, as well as their exposure to digital advertising campaigns.  In this way, we can leverage the 25,000+ data points we collect through panelist interviews to their browsing behaviors and online activities, providing a unique connection between brand engagement and the attributes, motivations and interests of our clients’ target audiences.”

“The GDPR was welcome validation of the approach we have always adopted, whereby respect for the consumer is central,” adds Smith. “Given that we have a direct relationship with the individuals who complete our surveys, we have an amazing opportunity to outline there and then what we want to do with their responses, and to gain their explicit consent for the use of cookies.

“The nature of survey-based market research also means that the consumers know which types of data they are providing, and can decline to answer certain questions if they so choose.”

So why does a company like Google — which has access to vast, global consumer data stores of its own, gathered from its own products and via a network of online tracking cookies and pixels — need GWI’s market intelligence?

“Many of our clients have their own proprietary sources of consumer data but turn to GlobalWebIndex because of the robust, global, independent view it offers on consumer behaviors,” Smith responds on this. “Our ability to give a 360-degree view on the consumer is particularly valuable, with our data providing a unique cross-device, cross-platform perspective.”

On the competition front, he names the likes of Nielsen, YouGov, Comscore, Kantar, GfK and Simmons. “We provide a global view which is consistent across countries, unlike some alternatives which operate different surveys at different times and then patch them together,” he argues.

He also touts the “incredibly rich view of the consumer’s digital life” the GWI consumer panel is able to generate, given the number of data points it’s gathering.

“We have this depth on all of our respondents, whereas some other sources will only have small pockets of data on each individual,” he claims. “We track behaviours and attitudes from the consumer’s perspective. There are no inferences, no modeling, no assumptions based on browsing. This is how the consumer acts and feels from their own point of view.

“We provide up-to-the minute data which tells you what people are doing in the here and now. Our quarterly releases will soon move to monthly, and from there we’ll develop a real-time version of our data set.”

“Our ability to re-interview our respondents is hugely important for clients, as it allows them to sync their bespoke questions with the 25,000 data points from the core study. It’s a hugely quick and efficient way to gain a rich understanding of your target audience,” he adds.

Read Full Article

Light Hacks

Light Hacks

Your Amazon Echo Now Has an Audio Equalizer

You can now change the audio equalizer settings on your Amazon Echo or other Alexa-enabled devices, and all using your voice. This means you can crank up the bass while playing dance music, or boost the treble when listening to podcasts in the morning.

The range of Amazon Echo devices is growing bigger with each passing month. So much so you may need help choosing which Echo device to buy. However, what they all have in common is the capacity to play audio content. Which is why the new EQ is a Godsend.

Amazon Adds New EQ Settings to the Echo

According to a post on the Amazon Developer blog, customers can “personalize the tone of the audio playing on their speakers, AV receivers, and sound bars by simply saying, ‘Alexa, crank up the bass,’ ‘Alexa, turn up treble,’ or ‘Alexa, set TV mode.'”

The EQ covers three bands, called Bass, Midrange, and Treble. All of which can be adjusted manually, with a range of plus or minus 6dBs. The new sound mode controls support Movie, Music, Night, Sport, and TV, which should cover most eventualities.

Anyone else whisper around their amazon echo because you think it’s listening to you all the time?

— Officer Jake Lukas (@officerjlukas) July 23, 2018

The devices include the Amazon Echo, Echo Dot, Echo Plus, Echo Show, and Echo Spot, as well as Alexa-enabled devices such as the Sonos Beam and Polk Command Bar. Amazon is opening up its API to other developers, meaning more devices will be added over time.

From Streaming Music to Awesome Podcasts

The equalizer and sound mode controls are currently only available to customers in the U.S. However, Amazon is promising to bring them to other English-speaking countries sometime in the future. As for everyone else, it’s anyone’s guess.

It’s actually rather surprising it has taken Amazon this long to add an equalizer to its Echo devices. After all, most people use their Echo to listen to something, whether that be streaming music using Alexa or listening to the best podcasts currently available.

Image Credit: Runran/Flickr

Read the full article: Your Amazon Echo Now Has an Audio Equalizer

Read Full Article

Segway’s whacky new roller shoes will cost $399

Did you know Segway is making a pair of self-balancing roller shoes? It is! The company has been tinkering with all sorts of new form factors since it was acquired by Ninebot in 2015, from half-sized Segways to kick scooters. Next up: inline… shoe… platform things.

Called the Segway Drift W1s, they sorta look like what would happen if you took a hoverboard (as in the trendy 2016 hoverboard-that-doesn’t-actually-hover “hover”board, not Marty McFly’s hoverboard), split it in two, and plopped one half under each foot.

It released a video demonstrating the shoes a few weeks back. Just watching it makes me feel like I’ve bruised my tailbone, because I’m clumsy as hell.

Pricing and availability was kept under wraps at the time, but the company has just released the details: a pair will cost you $399, and ship sometime in August. Oh, and they’ll come with a free helmet because you’ll probably want to wear a helmet.

A new product page also sheds some light on a few other previously undisclosed details: each unit will weigh about 7.7lbs, and top out at 7.5 miles per hour. Riding time “depends on riding style and terrain”, but the company estimates about 45 minutes of riding per charge.

I look forward to trying these then realizing I have absolutely no idea how to jump off and just riding forever into the sunset.

Read Full Article

Google wants Go to become the go-to language for writing cloud apps

The Google-incubated Go language is one of the fastest growing programming languages today, with about one million active developers using it worldwide. But the company believes it can still accelerate its growth, especially when it comes to its role in writing cloud applications. And to do this, the company today announced Go Cloud, a new open-source library and set of tools that makes it easier to build cloud apps with Go.

While Go is highly popular among developers, Google argues that the language was missing a standard library for interfacing with cloud services. Today, developers often have to essentially write their own libraries to use the features of each cloud, but organizations today want to be able to easily move their workloads between clouds.

What Go Cloud then gives these developers is a set of open generic cloud APIs for accessing blog storage, MySQL databases and runtime configuration, as well as an HTTP server with built-in logging, tracing and health checking. Right now, the focus is on AWS and Google Cloud Platform. Over time, Google plans to add more features to Go Cloud and support for more cloud providers (and those cloud providers can, of course, build their own support, too).

This, Google argues, allows developer teams to build applications that can easily run on any supported cloud without having to re-architect large parts of their applications.

As Google VP of developer relations Adam Seligman told me, the company hopes this move will kick off an explosion of libraries around Go — and, of course, that it will accelerate Go’s growth as a language for the cloud.

Read Full Article

Former Viki CEO Tammy Nam joins PicsArt as its first COO

PicsArt, the company behind the photo-editing app of the same name, has hired Tammy Nam as its first chief operating officer.

Nam was most recently the CEO of Viki, the Rakuten-acquired video streaming service, and before that served as a marketing executive at Viki, Scribd and Slide.

PicsArt said Nam will report to founder and CEO Hovhannes Avoyan, and that she will oversee all aspects of the business except for product and engineering.

“PicsArt has grown organically so far, but our next big opportunity is in directing this growth through the right market development, community engagement and revenue channels,” Avoyan said in the announcement. “In addition to her proven operational experience in both consumer advertising-and subscription-based businesses, Tammy adds deep bench strength in market, brand and community development — areas that will be critical for us moving forward.”

The company announced last year that it’s reaching 100 million monthly active users. Nam told me she was particularly impressed that it achieved that growth without significant marketing spend.

“I understand what it takes to grow quickly, but also thoughtfully,” she said. “Because of my background, the CEO and the board felt like I would be a great match to [help PicsArt] reach the next 200 million, the next 500 million users.”

Asked what thoughtful growth looks like for PicsArt, Nam said it means not just growing at any cost, but also considering at things like revenue and the different communities using the app. She said she’s trying to examine the company’s structure to ensure it can “maximize efficiencies towards these big goals.”

“It will continue to grow organically, but the branding, the user development will definitely evolve,” she added. “There’s a sea of companies that play in our space … How do you stand out? And how do you stay relevant?”

Nam also said that she’ll be looking at PicsArt’s opportunities for international growth. Not that the company has been neglecting the world beyond the United States — China is its fastest-growing market and already one of its top countries for revenue. (The company says it recently became profitable following the launch of its PicsArt Gold subscription.)

Nam suggested that PicsArt can move into new markets without competing with the dominant social media platforms, because it’s “agnostic” in terms of where users publish their edited photos.

“It’s completely lowered the barrier,” she said. “It used to be, you had to know Photoshop. Now it’s so easy to create professional-looking photos, images, and soon animations, videos, etc. Everyone is a creator.”

Read Full Article

Google Assistant’s news feeds are getting smarter

News is far and away the feature I use the most with Google Assistant. Every morning, I ask the Assistant “what’s in the news,” and it dutifully cycles through some pre-recorded news briefs from NPR, CNN and the like. It does the job, but it’s not much for specificity.

Google, however, is introducing tools to help developers target specific content based on queries. Per the example given in a new blog post, publishers can highlight a snippet of a story that will be read aloud when a user makes a request along the lines of “Hey Google, what’s the latest news on NASA?”Assistant will then read that portion aloud. The link to the full article is sent to the user’s mobile device and once done, Assistant will ask if they want another.

It’s interesting to watch companies like Google and Amazon play around with these news reads. It seems no one has quite figured out the ideal length for an audible news digests, but it appears to fall somewhere between a headline and full story. Or maybe it’s something more akin to bullet points, which the option to read on if the user wants more information.

Organizations like NPR and CNN do appear to have something of a head start, since a smart speaker briefing isn’t entirely dissimilar from get your information from cable news or public radio. Short, distilled snippets certainly seem like the way to go. As more people use the service and the AIs become more advanced, it will be easier to tailor that information to specific users.

At the very least, this should provide a way to further customize those feeds — not to mention giving Google even more insight into what its users are searching for. The feature will only be available for U.S. English speakers at launch. 

Read Full Article