Categories
AI

RISC-V grows open source processor membership 130% in 2021

Hear from CIOs, CTOs, and other C-level and senior execs on data and AI strategies at the Future of Work Summit this January 12, 2022. Learn more


RISC-V International said it has grown during the pandemic as its RISC-V open source processor membership popped 130% in 2021.

The nonprofit group’s membership has grown from a ragtag group of feisty academics to some of the biggest tech companies on earth like Google. Over the past decade, the group has groomed RISC-V into a viable alternative to proprietary Arm and Intel-based processors, and it appears that a lot of big companies and engineering geeks like what they see.

RISC-V chip revenues are expected to generate $400 million in 2021 and reach $1 billion in revenue by 2024, according to a prediction made this week by accounting and consulting firm Deloitte. The firm said the ripples of RISC-V could turn into the waves of the future. Calista Redmond, CEO of RISC-V International, said in an interview with VentureBeat that there are 2,4278 members in the group now, up 130% since the start of the year, and 292 companies, up 27.5% this year.

“We’ve increased and grown and we have seen a deeper investment coming in from around the world, as you see that the reflection in our board composition, which is both premier members who are coming in at our highest level of membership, as well as elected representatives of various groups,” Redmond said. “That deeper investment is a reflection of them bringing RISC-V across their portfolio of products rather than just isolated to a few projects.”

The RISC-V membership will be gathering in San Francisco both in-person and online at an event anchored at the Moscone convention center. The members are announcing today that they have ratified 15 new specifications — representing more than 40 extensions to the hardware architecture — for the RISC-V instruction set architecture, which anyone can use for free.

Calista Redmond, CEO of the RISC-V Foundation.

Above: Calista Redmond, CEO of the RISC-V Foundation.

Image Credit: RISC-V Foundation

Redmond said one of the benefits of RISC-V is that it is sanction-free. As an open source platform, RISC-V is not affected by export restrictions. This makes it appealing to companies, especially in China, that have been affected or fear being affected by those restrictions, Deloitte said. Redmond said that the nonprofit had to reboot its entire membership base as it transferred its headquarters from the U.S. to Switzerland in order to erase any doubt that it was independent of geographic borders.

Deloitte said that companies are planning on using it for different storage, graphics, and machine-learning applications. Even Intel’s foundry services division is partnering with RISC-V player SiFive. Arm argues that it has more features and has more support options for developers. Since Arm is based in the United Kingdom and Intel in the U.S., Chinese manufacturers worry that they could lose access to the architectures if trade friction heats up. Nvidia is hoping regulators will approve its plan to pay $40 billion to buy Arm.

Redmond said there are a lot of Chinese members, but overall RISC-V’s base is about a third North America, a third European, and a third Asian.

“We’ve always been global. There is nothing that changed at all in the rules, regulations, or global constructs that we participate in,” Redmond said. “Our move was primarily just to address any concerns that the landscape could change.”

Redmond said that designers don’t have to worry about constraints on what they do and that gives them freedom for innovation.

Deloitte also said that startups care about the royalty-free open source architecture. In the three years between 2020 and 2022, venture capitalists (VCs) will invest about $22 billion into startup chip companies of all kinds, Deloitte said. A million-dollar license fee may not matter to one of the world’s largest smartphone companies, but it does matter for a startup that has relatively little cash and a monthly burn rate, Deloitte said.

The served addressable market (SAM) for RISC-V in automotive alone was 4 million cores in 2020, forecast to rise to 150 million cores in 2022, and to 2.9 billion cores by 2025.

New specifications

Above: RISC-V software is expected to grow dramatically.

Image Credit: Tractica

Mark Himelstein, chief technology officer of RISC-V International, said in an interview with VentureBeat that the specifications cover vector, scalar cryptography, and hypervisor features that will keep extending the reach of RISC-V processors into new markets. Developers will find it easier to create RISC-V applications for artificial intelligence (AI) and machine learning (ML), the Internet of Things (IoT), connected and autonomous cars, data centers, and more, he said.

“The development of these specifications really showcased the incredible benefits of open collaboration across companies and geographies as members worked together to develop novel approaches for the latest computing requirements,” said Krste Asanović, chair of the RISC-V International, in a statement.

Redmond and Himelstein said that RISC-V’s advantages include that designs based on it are easy to modify. As such, they can offer greater flexibility than traditional chip designs.

“All 15 of those specifications have been ratified by the board. They all have passed acceptance criteria. And we’re very excited about that,” Himelstein said. “And we have another bunch on deck.”

The RISC-V Vector specification will help accelerate the computation of data intensive operations like ML inference for audio, vision, and voice processing. With RISC-V Vector, developers can process complex data arrays and scalar operations quickly and with low latency. The simplicity and flexibility of Vector allows companies to easily customize RISC-V solutions for a wide variety of edge computing applications from consumer IoT devices to industrial ML applications.

“The new RISC-V Vector specification will change the way people think about vector designs,” said Dave Ditzel, executive chairman of Esperanto Technologies, in a statement. “With just over 100 instructions, the extension offers a simple and elegant approach to efficiently process the latest machine learning algorithms.”

The RISC-V Hypervisor specification virtualizes supervisor-level architecture to efficiently host guest operating systems atop a type-1 or type-2 hypervisor. Virtual machine implementations require the RISC-V Hypervisor specification. The Hypervisor specification will help drive RISC-V adoption in cloud and embedded applications where virtualization is critical, such as in data centers, automotive applications, and industrial control applications. The RISC-V community has ported KVM and other open source virtual machines on top of simulators using the new specification.

The RISC-V Scalar Cryptography specification enables the acceleration of cryptographic workloads for small footprint deployments. These extensions significantly lower the barrier to entry for secure and efficient accelerated cryptography in IoT and embedded devices.

“The RISC-V Scalar Cryptography extensions allow for implementing standard cryptographic hash and block cipher algorithms that are an order of magnitude faster than using standard instructions in some cases. With RISC-V’s transparent and open approach, anyone can efficiently implement critical cryptographic algorithms in any class of CPU,” said Ben Marshall, cryptographic hardware engineer at PQShield and member of the RISC-V Technical Steering Committee, in a statement. “In addition to the performance benefits, these new extensions are very cheap to implement so companies can integrate popular cryptography algorithms in even the smallest connected devices.”

VentureBeat

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

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

  • up-to-date information on the subjects of interest to you
  • our newsletters
  • gated thought-leader content and discounted access to our prized events, such as Transform 2021: Learn More
  • networking features, and more

Become a member

Repost: Original Source and Author Link

Categories
AI

Membership inference attacks detect data used to train machine learning models

Join Transform 2021 this July 12-16. Register for the AI event of the year.


One of the wonders of machine learning is that it turns any kind of data into mathematical equations. Once you train a machine learning model on training examples—whether it’s on images, audio, raw text, or tabular data—what you get is a set of numerical parameters. In most cases, the model no longer needs the training dataset and uses the tuned parameters to map new and unseen examples to categories or value predictions.

You can then discard the training data and publish the model on GitHub or run it on your own servers without worrying about storing or distributing sensitive information contained in the training dataset.

But a type of attack called “membership inference” makes it possible to detect the data used to train a machine learning model. In many cases, the attackers can stage membership inference attacks without having access to the machine learning model’s parameters and just by observing its output. Membership inference can cause security and privacy concerns in cases where the target model has been trained on sensitive information.

From data to parameters

deep neural network AI

Above: Deep neural networks use multiple layers of parameters to map input data to outputs

Each machine learning model has a set of “learned parameters,” whose number and relations vary depending on the type of algorithm and architecture used. For instance, simple regression algorithms use a series of parameters that directly map input features to the model’s output. Neural networks, on the other hand, use complex layers of parameters that process input and pass them on to each other before reaching the final layer.

But regardless of the type of algorithm you choose, all machine learning models go through a similar process during training. They start with random parameter values and gradually tune them to the training data. Supervised machine learning algorithms, such as those used in classifying images or detecting spam, tune their parameters to map inputs to expected outcomes.

For example, say you’re training a deep learning model to classify images into five different categories. The model might be composed of a set of convolutional layers that extract the visual features of the image and a set of dense layers that translate the features of each image into confidence scores for each class.

The model’s output will be a set of values that represent the probability that an image belongs to each of the classes. You can assume that the image belongs to the class with the highest probability. For instance, an output might look like this:

Cat: 0.90
Dog: 0.05
Fish: 0.01
Tree: 0.01
Boat: 0.01

Before training, the model will provide incorrect outputs because its parameters have random values. You train it by providing it with a collection of images along with their corresponding classes. During training, the model gradually tunes the parameters so that its output confidence score becomes as close as possible to the labels of the training images.

Basically, the model encodes the visual features of each type of image into its parameters.

Membership inference attacks

A good machine learning model is one that not only classifies its training data but generalizes its capabilities to examples it hasn’t seen before. This goal can be achieved with the right architecture and enough training data.

But in general, machine learning models tend to perform better on their training data. For example, going back to the example above, if you mix your training data with a bunch of new images and run them through your neural network, you’ll see that the confidence scores it provides on the training examples will be higher than those of the images it hasn’t seen before.

training examples vs new examples

Above: Machine learning models perform better on training examples as opposed to unseen examples

Membership inference attacks take advantage of this property to discover or reconstruct the examples used to train the machine learning model. This could have privacy ramifications for the people whose data records were used to train the model.

In membership inference attacks, the adversary does not necessarily need to have knowledge about the inner parameters of the target machine learning model. Instead, the attacker only knows the model’s algorithm and architecture (e.g., SVM, neural network, etc.) or the service used to create the model.

With the growth of machine learning as a service (MaaS) offerings from large tech companies such as Google and Amazon, many developers are compelled to use them instead of building their models from scratch. The advantage of these services is that they abstract many of the complexities and requirement of machine learning, such as choosing the right architecture, tuning hyperparameters (learning rate, batch size, number of epochs, regularization, loss function, etc.), and setting up the computational infrastructure needed to optimize the training process. The developer only needs to set up a new model and provide it with training data. The service does the rest.

The tradeoff is that if the attackers know which service the victim used, they can use the same service to create a membership inference attack model.

In fact, at the 2017 IEEE Symposium on Security and Privacy, researchers at Cornell University proposed a membership inference attack technique that worked on all major cloud-based machine learning services.

In this technique, an attacker creates random records for a target machine learning model served on a cloud service. The attacker feeds each record into the model. Based on the confidence score the model returns, the attacker tunes the record’s features and reruns it by the model. The process continues until the model reaches a very high confidence score. At this point, the record is identical or very similar to one of the examples used to train the model.

membership inference attack models

Above: Membership inference attacks observe the behavior of a target machine learning model and predict examples that were used to train it.

After gathering enough high confidence records, the attacker uses the dataset to train a set of “shadow models” to predict whether a data record was part of the target model’s training data. This creates an ensemble of models that can train a membership inference attack model. The final model can then predict whether a data record was included in the training dataset of the target machine learning model.

The researchers found that this attack was successful on many different machine learning services and architectures. Their findings show that a well-trained attack model can also tell the difference between training dataset members and non-members that receive a high confidence score from the target machine learning model.

The limits of membership inference

Membership inference attacks are not successful on all kinds of machine learning tasks. To create an efficient attack model, the adversary must be able to explore the feature space. For example, if a machine learning model is performing complicated image classification (multiple classes) on high-resolution photos, the costs of creating training examples for the membership inference attack will be prohibitive.

But in the case of models that work on tabular data such as financial and health information, a well-designed attack might be able to extract sensitive information, such as associations between patients and diseases or financial records of target people.

overfitting vs underfitting

Above: Overfitted models perform well on training examples but poorly on unseen examples.

Membership inference is also highly associated with “overfitting,” an artifact of poor machine learning design and training. An overfitted model performs well on its training examples but poorly on novel data. Two reasons for overfitting are having too few training examples or running the training process for too many epochs.

The more overfitted a machine learning model is, the easier it will be for an adversary to stage membership inference attacks against it. Therefore, a machine model that generalizes well on unseen examples is also more secure against membership inference.

This story originally appeared on Bdtechtalks.com. Copyright 2021

VentureBeat

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

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

  • up-to-date information on the subjects of interest to you
  • our newsletters
  • gated thought-leader content and discounted access to our prized events, such as Transform 2021: Learn More
  • networking features, and more

Become a member

Repost: Original Source and Author Link

Categories
Tech News

Best Buy paid membership program is its Amazon Prime response

Amazon really turned the retail industry on its head and even until now big shot competitors from brick and mortar businesses are still trying to catch up to it. One of Amazon’s biggest weapons is Prime membership and its perks and rivals are still trying their hand at offering their customers something similar. After Walmart’s own Walmart+, Best Buy is now testing its own paid subscription program but with a more “human” spin to set it apart from Amazon Prime.

At this point in time, it’s almost futile to compete with Amazon Prime’s perks when it comes to pricing and the delivery of goods. Not that Walmart or Best Buy are incapable of matching whatever Amazon offers, just that they might not be able to offer more than that. Best Buy, however, seems to think it can by adding a human element to its premium service.

The pilot membership program christened Best Buy Beta does have the usual exclusive discounts, fast and free shipping, and other shopping perks you might expect from such a retail-oriented subscription service. Beyond that, however, it is also offering services like exclusive concierge support, free home installation on most products, and unlimited tech support from its famous Geek Squad.

It is questionable whether those perks can really drive Best Buy Beta’s success but it does underscore something that is often taken for granted. Not everyone buying new products, especially technological and electronic ones, might have the know-how to navigate the installation and use of these devices. What Best Buy Beta practically offers isn’t just convenience but some assurance of human support as well.

That, however, does come at a steeper price than others, with the premium membership costing $199.99 per year. It is currently available only at select stores in Iowa, Oklahoma, and eastern Pennsylvania but is planned to expand to Minnesota, North Carolina, and Tennessee this month.

Repost: Original Source and Author Link

Categories
Game

PS Plus Deal Gives You Whole Year’s Membership for Only $30

Digital Trends may earn a commission when you buy through links on our site.

While the PlayStation 5 is currently the target for most gamers looking for PlayStation deals, there are still a lot of reasons to buy or stay with the PlayStation 4. However, whether you’re moving up to PlayStation 5 or sticking with the PlayStation 4, you will need a PlayStation Plus subscription to maximize the capabilities of your console. If your budget’s already allocated for upcoming games, the good news is that there are PS Plus deals that will let you access membership benefits at a cheaper cost. One of the sources for such offers is CDKey, which is selling the 3-month PS Plus membership for $14, for 50% off its original price of $28, and the 1-year PS Plus membership for $30, also 50% off its original price of $60.

3-month membership — $14, was $28

1-year membership — $30, was $60

The main purpose of a PlayStation Plus subscription is to activate the online multiplayer modes of the games that you purchase. While it’s possible to play most of these games without PS Plus, there’s simply no substitute for the online multiplayer experience in titles like Call of Duty: Black Ops Cold War, NBA 2K21, Marvel’s Avengers, and Overwatch, as well as Red Dead Redemption 2‘s Red Dead Online and Grand Theft Auto V‘s Grand Theft Auto Online.

Another huge bonus that comes with a PlayStation Plus membership is that you gain access to free game downloads that change every month. The free games for March are the modernized RPG Final Fantasy VII Remake, the third-person survival shooter Remnant: From the Ashes, and the vehicle-based arena battler Destruction AllStars. PlayStation 5 owners will also be able to download first-person puzzler Maquette, while those who own the PlayStation VR system can enjoy space adventure Farpoint.

PlayStation Plus members will also enjoy exclusive discounts to games, add-ons, and pre-orders on the PlayStation Store, as well as gain access to 100GB of cloud storage where your saves may be retrieved, in case anything happens to your console.

Whether you’re among the lucky ones who have secured a PlayStation 5, or part of the crowd who are still happy with the PlayStation 4, PlayStation Plus is a requirement to make the most out of your console. The subscription doesn’t have to dig in too much into your gaming budget though. CDKeys is offering a $14 discount on the 3-month PS Plus membership, lowering its price to $14 from $28, and a $30 discount on the 1-year PS Plus membership, bringing its price down to $30 from $60. It’s unclear when the deal will end, so if you want to start playing online and downloading free games as soon as possible, you should click the Buy Now button immediately.

3-month membership — $14, was $28

1-year membership — $30, was $60

We strive to help our readers find the best deals on quality products and services, and we choose what we cover carefully and independently. The prices, details, and availability of the products and deals in this post may be subject to change at anytime. Be sure to check that they are still in effect before making a purchase.

Digital Trends may earn commission on products purchased through our links, which supports the work we do for our readers.

Editors’ Choice




Repost: Original Source and Author Link

Categories
Game

PS Plus Sale Will Give You a Whole Year’s Membership for $30

Digital Trends may earn a commission when you buy through links on our site.

Whether you’ve upgraded to the PlayStation 5 or you’re sticking with the massive games library of the PlayStation 4, subscribing to a PlayStation Plus membership will make you appreciate your console more. However, if you want to spend most of your budget on video game deals, the good news is that there are some PS Plus deals out there that will let you sign up to the service for a cheaper price. One of the reliable source for PS Plus discounts is CDKeys, which is offering a $14 discount for a 3-month PS Plus membership, bringing its price down to $14 from its original price of $28, and a $30 discount for a 1-year membership, bringing its price down to $30 from its original price of $60.

3-month PS Plus membership – $14, was $28;

1-year PS Plus membership – $30, was $60;

The primary purpose of a PlayStation Plus subscription is to enable online multiplayer modes, which is part of the full experience for first-person shooters like Call of Duty: Black Ops Cold War, co-op adventures like Monster Hunter: World, and sports simulators like NBA 2K21. PS Plus also lets you jump into Red Dead Redemption 2‘s Red Dead Online and Grand Theft Auto V‘s Grand Theft Auto Online, for deeper immersion in these open-world titles.

An active PS Plus subscription will also let you download free games each month, which you can keep playing as long as your membership remains active. The free PS Plus games for March are the critically acclaimed action RPG Final Fantasy VII Remake and third-person survival shooter Remnant: From the Ashes for the PlayStation 4, first-person puzzle game Maquette for the PlayStation 5, and space adventure Farpoint for the PlayStation VR.

If you own a PlayStation 5, a PlayStation 4, or both, it’s highly recommended that you also sign up for a PlayStation Plus membership in order to fully maximize the capabilities of Sony’s consoles. You don’t have to spend much on a subscription with CDKeys’ offers that slash 50% off their price, lowering the price of the 3-month PS Plus membership to $14 from its original price of $28 and the 1-year PS Plus membership to $30 from its original price of $60. To take advantage of this month’s free PS Plus games and to start playing online multiplayer modes, you should click that Buy Now button as soon as you can.

3-month PS Plus membership – $14, was $28;

1-year PS Plus membership – $30, was $60;

We strive to help our readers find the best deals on quality products and services, and we choose what we cover carefully and independently. The prices, details, and availability of the products and deals in this post may be subject to change at anytime. Be sure to check that they are still in effect before making a purchase.

Digital Trends may earn commission on products purchased through our links, which supports the work we do for our readers.

Editors’ Choice




Repost: Original Source and Author Link

Categories
Game

PS Plus One-Year Membership Drops to Lowest Price Ever

Digital Trends may earn a commission when you buy through links on our site.

If you’ve been waiting to elevate your PlayStation gaming experience with online play, now’s the time to grab a three-month or one-year membership at the best discount we’ve ever seen. Earlier this month, CDKeys offered a 50% discount on the one-year membership, beating Black Friday and Cyber Monday pricing, and the retailer has outdone itself again with the current sale. You can now get a three-month membership for $11, down from $28, or a one-year membership for $28, down from $60. Save 53% and score the cheapest price yet on a PS Plus membership.

Buy One-Year Membership

Buy Three-Month Membership

You can compete against friends or strangers in your favorite games with online multiplayer access, and the benefits don’t end there. You’ll also get two free games each month, from new indie games to blockbuster classics, which are yours to keep for the duration of your membership. In addition, you’ll get exclusive discounts on add-ons, new games, pre-orders, and more.

PS Plus members get access to free extras not typically included in free games such as Fortnite, like weapons, skins, and cosmetics. And you can save your progress and easily pick up where you left off with 100GB free cloud storage. Plus, your games update automatically in the cloud, so they’re ready when you are. Your membership will work on all PS3, PS4, and PS Vita systems. Check out the best PS4 games you can play right now if you’re unsure where to start, or simply check out this month’s free games to get started at no additional cost.

Even if you already have a PS Plus membership, now would be a great time to take advantage of this awesome discount and extend your membership another three months or a year. You can check out other PS Plus deals as well, but this is the best discount around. Delivery is instant for both options, so if you’re new to PS Plus, you’ll be able to get started right away. You can save $32 on the one-year membership at CDKeys, dropping your total from $60 to just $28 for 12 full months of access, or you can save $17 on the three-month membership, which is reduced from the list price of $28 to just $11. Since this is the best deal we’ve seen yet on a PS Plus membership, you’ll want to act fast to take advantage of it.

Buy One-Year Membership

Buy Three-Month Membership

We strive to help our readers find the best deals on quality products and services, and we choose what we cover carefully and independently. The prices, details, and availability of the products and deals in this post may be subject to change at anytime. Be sure to check that they are still in effect before making a purchase.

Digital Trends may earn commission on products purchased through our links, which supports the work we do for our readers.

Editors’ Choice




Repost: Original Source and Author Link

Categories
Tech News

A Stone River membership gives you 800 tech courses and 4,800 hours of training for under $100

TLDR: A Stone River eLearning lifetime membership offers access to over 800 courses for learning everything tech, from coding to web design to game development and beyond.

There’s no industry that moves faster than tech. The newest and hottest thing one week becomes old hat in need of a 2.0 version the next. That puts tech professionals in the anxious position of knowing there’s always more they need to know to stay on top of the latest innovations and techniques.

Even the most connected and inspired tech practitioners need a reservoir of knowledge to fall back on when they need to add a new skill to their toolkit. Stone River eLearning ($89, over 90 percent off, from TNW Deals) can be that backstop, offering detailed online training in dozens of tech disciplines to keep you up to date on your current skills, or even open up a whole new area to explore.

Students are hard-pressed to find anything this online academy doesn’t cover with loads of potential coursework. The Stone River catalog is vast, home to over 800 courses covering more than 4,800 hours of intensive instruction.

If you’re invested in learning to create video games, there are literally two dozen courses in learning everything from Unity and 3D gaming to mobile development and game hacking to turn out a finished game in just one hour. Meanwhile, another two dozen courses delve into the DevOps field, including training in Jenkins, Docker, AWS, Kubernetes, and more.

Even if you just want to round out your general programming, web design, or cybersecurity knowledge, you’ll find Stone River courses explaining it all, like Bootstrap, Java, Python, MySQL or node.js. And that’s all just scratching the surface of the depth of training available through a Stone River membership.

With over 500,000 students already, Stone River offers other VIP perks like easily obtainable ebooks, personal guidance on areas to learn, and even a wealth of certification exams, many of which will usually cost more than $50 each to take.

Right now, you can move your career forward or just pick up a fun new hobby with a Stone River eLearning membership. A lifetime of course access is regularly over $11,000, but as part of this collection, you’re a member forever for only $89.

Prices are subject to change.

Repost: Original Source and Author Link