Windows 10: Samsung to Strengthen NPU Capabilities for Future AI Applications

Brink · Jun 18, 2019

Samsung Electronics, a world leader in advanced semiconductor technology, today announced that it will strengthen its neural processing unit (NPU) capabilities to further extend the reach of its artificial intelligence (AI) solutions.

In line with its focus on next-generation NPU technologies, Samsung plans to create over 2,000 related jobs worldwide by 2030, which is about 10 times the current headcount. The company will also expand upon its existing collaboration with globally distinguished research institutes and universities, and support the nurturing of future talent in the field of AI, including deep learning and neural processing.

“For the coming age of AI, Samsung is committed to delivering industry-leading advancements brought to life by our NPU technologies,” said Inyup Kang, president of System LSI Business at Samsung Electronics. “As we leverage our differentiated technology, close partnerships with global institutes and active investment in top talent, we are excited to take future AI processing capabilities to the next level.”

Samsung introduced its first NPU in the company’s premium mobile processor, the Exynos 9820, last year and plans to continue offering advanced on-device AI features for high-performance mobile chips. Applications for Samsung’s NPUs will expand into areas such as automotive processors that power in-vehicle infotainment (IVI) and advanced driver assistance systems (ADAS), as well as next-generation datacenters optimized for big data processing.

The System LSI Business and Samsung Advanced Institute of Technology — Samsung’s R&D arm — together plan to extend and evolve the company’s current NPU research into novel AI hardware technologies such as neuromorphic processors that aim to operate at the level of a human brain.
Click to expand...

Source: Samsung Electronics to Strengthen its Neural Processing Capabilities for Future AI Applications – Samsung Global Newsroom

:)

Brink · Jun 18, 2019

Securing the future of AI and machine learning at Microsoft

Artificial intelligence (AI) and machine learning are making a big impact on how people work, socialize, and live their lives. As consumption of products and services built around AI and machine learning increases, specialized actions must be undertaken to safeguard not only your customers and their data, but also to protect your AI and algorithms from abuse, trolling, and extraction.

We are pleased to announce the release of a research paper, Securing the Future of Artificial Intelligence and Machine Learning at Microsoft, focused on net-new security engineering challenges in the AI and machine learning space, with a strong focus on protecting algorithms, data, and services. This content was developed in partnership with Microsoft’s AI and Research group. It’s referenced in The Future Computed: Artificial Intelligence and its role in society by Brad Smith and Harry Shum, as well as cited in the Responsible bots: 10 guidelines for developers of conversational AI.

This document focuses entirely on security engineering issues unique to the AI and machine learning space, but due to the expansive nature of the InfoSec domain, it’s understood that issues and findings discussed here will overlap to a degree with the domains of privacy and ethics. As this document highlights challenges of strategic importance to the tech industry, the target audience for this document is security engineering leadership industry-wide.

Our early findings suggest that:

1. Secure development and operations foundations must incorporate the concepts of Resilience and Discretion when protecting AI and the data under its control.

AI-specific pivots are required in many traditional security domains such as Authentication, Authorization, Input Validation, and Denial of Service mitigation.

Without investments in these areas, AI/machine learning services will continue to fight an uphill battle against adversaries of all skill levels.

2. Machine learning models are largely unable to discern between malicious input and benign anomalous data. A significant source of training data is derived from un-curated, unmoderated public datasets that may be open to third-party contributions.

Attackers don’t need to compromise datasets when they are free to contribute to them. Such dataset poisoning attacks can go unnoticed while model performance inexplicably degrades.

Over time, low-confidence malicious data becomes high-confidence trusted data, provided that the data structure/formatting remains correct and the quantity of malicious data points is sufficiently high.

3. Given the great number of layers of hidden classifiers/neurons that can be leveraged in a deep learning model, too much trust is placed on the output of AI/machine learning decision-making processes and algorithms without a critical understanding of how these decisions were reached.

AI/machine learning is increasingly used in support of high-value decision-making processes in medicine and other industries where the wrong decision may result in serious injury or death.

AI must have built-in forensic capabilities. This enables enterprises to provide customers with transparency and accountability of their AI, ensuring its actions are not only verifiably correct but also legally defensible.

When combined with data provenance/lineage tools, these capabilities can also function as an early form of “AI intrusion detection,” allowing engineers to determine the exact point in time that a decision was made by a classifier, what data influenced it, and whether or not that data was trustworthy.

Our goal is to bring awareness and energy to the issues highlighted in this paper while driving new research investigations and product security investments across Microsoft. Read the Securing the Future of Artificial Intelligence and Machine Learning at Microsoft paper to learn more.
Click to expand...

Source: Securing the future of AI and machine learning at Microsoft - Microsoft Secure

Brink · Jun 18, 2019

How future policy and regulations will challenge AI

I recently wrote about how radical the incorporation of artificial intelligence (AI) to cybersecurity will be. Technological revolutions are however frequently not as rapid as we think. We tend to see specific moments, from Sputnik in 1957 to the iPhone in 2007, and call them “game changing” – without appreciating the intervening stages of innovation, implementation and regulation, which ultimately result in that breakthrough moment. What can we therefore expect from this iterative and less eye-catching part of AI’s development, looking not just at the technological progress, but its interaction with national policy-making process?

I can see two overlapping, but distinct, perspectives. The first relates to the reality that information and communication technology (ICT) and its applications develop faster than laws. In recent years, examples of social media and/or ride hailing apps have seen this translate into the following regulatory experience:

Innovation: R&D processes arrive at one or many practical options for a technology;

Implementation: These options are applied in the real world, are refined through experience, and begin to spread through major global markets;

Regulation: Governments intervene to defend the status quo or to respond to new categories of problem, e.g. cross-border data flows;

Unanticipated consequences: Policy and technology’s interaction inadvertently harms one or both, e.g. the Wassenaar’s impact on cybersecurity R&D.

AI could follow a similar path. However, unlike e-commerce or the shared economy (but like nanotechnology or genetic engineering) AI actively scares people, so early regulatory interventions are likely. For example, a limited focus on using AI in certain sectors, e.g. defense or pharmaceuticals, might be positioned as more easily managed and controlled than AI’s general application. However, could such a limit really be imposed, particularly in the light of potential for transformative creative leaps that AI seems to promise? I say that would be unlikely – resulting in yet more controls. Leaving aside the fourth stage of unknown unknowns of unanticipated consequences, the third phase, i.e. regulation, would almost inevitably run into trouble of its own by virtue to having to legally define something as unprecedented and mutable as AI. It seems to me, therefore, that even the basic phases of AI’s interaction with regulation could be fraught with problems for innovators, implementers and regulators.

The second, more AI-specific perspective is driven by the way its capabilities will emerge, which I feel will break down into three basic stages:

Distinction: Creation of smarter sensors;

Direction: Automation of human-initiated decision-making;

Delegation: Enablement of entirely independent decision-making.

Smarter sensors will come in various forms, not least as part of the Internet of Things (IoT), and their aggregated data will have implications for privacy. 20th century “dumb lenses” are already being connected to systems that can pick out number plates or human faces but truly smart sensors could know almost anything about us, from what is in our fridge and on our grocery list, to where we are going and whom we will meet. It is this aggregated, networked aspect of smarter sensors that will be at the core of the first AI challenge for policy-makers. As they become discriminating enough to anticipate what we might do next, e.g. in order to offer us useful information ahead of time, they create an inadvertent panopticon that the unscrupulous and actively criminal can exploit.

Moving past this challenge, AI will become able to support and enhance human decision-making. Human input will still be essential but it might be as limited as a “go/no go” on an AI-generated proposal. From a legal perspective, mens rea or scope of liability might not be wholly thrown into confusion, as a human decision-maker remains. Narrow applications in certain highly technical areas, e.g. medicine or engineering, might be practical but day-to-day users could be flummoxed if every choice had unreadable but legally essential Terms & Conditions. The policy-making response may be to use tort/liability law, obligatory insurance for AI providers/users, or new risk management systems to hedge the downside of AI-enhanced decision-making without losing the full utility of the technology.

Once decision-making is possible without human input, we begin to enter the realm of speculation. However, it is important to remember that there are already high-frequency trading (HFT) systems in financial markets that operate independent of direct human oversight, following algorithmic instructions. The suggested linkages between “flash crash” events and HFT highlight, nonetheless, the problems policy-makers and regulators will face. It may be hard to foresee what even a “limited” AI might do in certain circumstances, and the ex-ante legal liability controls mentioned above may seem insufficient to policy-makers should a system get out of control, either in the narrow sense of being out of the control of those people legally responsible for it, or in the general sense of it being out of control of anybody.

These three stages would suggest significant challenges for policy-makers, with existing legal processes losing their applicability as AI moves further away from direct human responsibility. The law is, however adaptable, and solutions could emerge. In extremis we might, for example, be willing to add to the concept of “corporate persons” with a concept of “artificial persons”. Would any of us feel safer if we could assign legal liability to the AIs themselves and then sue them as we do corporations and businesses? Maybe.

In summary then, the true challenges for AI’s development may not exist solely in the big ticket moments of beating chess masters or passing Turing Tests. Instead, there will be any number of roadblocks caused by the needs of regulatory and policy processes systems still rooted in the 19th and 20th centuries. And, odd though this may sound from a technologist like me, that delay might be a good thing, given the potential transformative power of AI.
Click to expand...

Source: How future policy and regulations will challenge AI | Microsoft Secure Blog

ne6togadno · Jun 18, 2019

OFFICIAL Dragon's Dogma: Dark Arisen (Review)

there is something wrong with suggested for you list on npu page. clicking on picture of dd:da review from suggested for you list leads to perormance analysis instead of review. hovering mouse over it shows in status bar correct link to review but clicking on picture reloads analysis.
links from reviews&articles page works fine.

edit: link from home page of npu sends you to analysis instead of review as well.

Windows 10: Samsung to Strengthen NPU Capabilities for Future AI Applications

Samsung to Strengthen NPU Capabilities for Future AI Applications

Samsung to Strengthen NPU Capabilities for Future AI Applications

Samsung to Strengthen NPU Capabilities for Future AI Applications

Samsung to Strengthen NPU Capabilities for Future AI Applications - Similar Threads - Samsung Strengthen NPU

No NPU activity

No NPU activity

After installing Ryzen AI 1.2, the camera on Windows 11 23H2 is unable to utilize the NPU.

1st gen NPU users won't get AI features on 24H2?

1st gen NPU users won't get AI features on 24H2?

How AI could shape the future of journalism

Securing the future of AI and machine learning at Microsoft

Advancing the future of society with AI and the intelligent edge

How future policy and regulations will challenge AI