Our society increasingly benefits from Artificial Intelligence (AI). Unfortunately, more and more evidence shows that AI is also used for offensive purposes. Prior works have revealed various examples of use cases in which the deployment of AI can lead to violation of security and privacy objectives. No extant work, however, has been able to draw a holistic picture of the offensive potential of AI. In this SoK paper we seek to lay the ground for a systematic analysis of the heterogeneous capabilities of offensive AI. In particular we (i) account for AI risks to both humans and systems while (ii) consolidating and distilling knowledge from academic literature, expert opinions, industrial venues, as well as laypeople -- all of which being valuable sources of information on offensive AI. To enable alignment of such diverse sources of knowledge, we devise a common set of criteria reflecting essential technological factors related to offensive AI. With the help of such criteria, we systematically analyze: 95 research papers; 38 InfoSec briefings (from, e.g., BlackHat); the responses of a user study (N=549) entailing individuals with diverse backgrounds and expertise; and the opinion of 12 experts. Our contributions not only reveal concerning ways (some of which overlooked by prior work) in which AI can be offensively used today, but also represent a foothold to address this threat in the years to come.

Summary of paper: This is a technical paper on supervised sparse coding. It defines a generalized form of Lasso problem that equates some input variables x (e.g. the input image) to some hidden variables y (e.g. the underlying, analyzed or denoised form of the image) in least squares via two arbitrary weight matrices -- M1*x-M2*y 2 -- adding an L1 penalty on a linear mapping Omega*y and an optional L2 penalty on y itself. This formulation is general enough to cover both analysis and synthesis (generative) problems. The paper solves this system for y(x) via a proximal iteration on auxilliary variables z Omega*y -- an ADMM style method -- and it also offers fast approximation of y(x) via a network that contains an unwound, truncated ADMM with re-learned parameters along the lines of Gregor & LeCun [11]. Finally, it proposes supervised parameter (mapping matrix) learning using stochastic gradient descent over an arbitrary problem-specific loss function on y(x), and to this end it derives some explicit formulae for cost gradients in terms of sign(Omega*y) and the matrices and auxilliary vectors involved. The method is illustrated on an image super-resolution problem and a polyphonic music transcription one.

Homestay broker Airbnb found that the key to creating diversity with its machine learning algorithms is to have one neural network for standard learning and another to specifically diversify the results. Diversification means, in this sense, increasing the variety of options that a user sees when looking to book a stay. This concept took years to develop because it went directly against the company engineer's original core belief, that the probability of someone booking a listing could be determined independently of the listings themselves. Airbnb published a technical paper and a blog post detailing the quiet shortcomings of that approach: that a single neural network produces precise but homogeneous set of results. To increase diversity, Airbnb engineers underwent the process of creating and iterating on additional neural networks.

Realizing today's cloud-level artificial intelligence functionalities directly on devices distributed at the edge of the internet calls for edge hardware capable of processing multiple modalities of sensory data (e.g. video, audio) at unprecedented energy-efficiency. AI hardware architectures today cannot meet the demand due to a fundamental "memory wall": data movement between separate compute and memory units consumes large energy and incurs long latency. Resistive random-access memory (RRAM) based compute-in-memory (CIM) architectures promise to bring orders of magnitude energy-efficiency improvement by performing computation directly within memory. However, conventional approaches to CIM hardware design limit its functional flexibility necessary for processing diverse AI workloads, and must overcome hardware imperfections that degrade inference accuracy. Such trade-offs between efficiency, versatility and accuracy cannot be addressed by isolated improvements on any single level of the design. By co-optimizing across all hierarchies of the design from algorithms and architecture to circuits and devices, we present NeuRRAM - the first multimodal edge AI chip using RRAM CIM to simultaneously deliver a high degree of versatility for diverse model architectures, record energy-efficiency $5\times$ - $8\times$ better than prior art across various computational bit-precisions, and inference accuracy comparable to software models with 4-bit weights on all measured standard AI benchmarks including accuracy of 99.0% on MNIST and 85.7% on CIFAR-10 image classification, 84.7% accuracy on Google speech command recognition, and a 70% reduction in image reconstruction error on a Bayesian image recovery task. This work paves a way towards building highly efficient and reconfigurable edge AI hardware platforms for the more demanding and heterogeneous AI applications of the future.

Overview: Machine learning is often portrayed as a value-neutral endeavor; even when that is not the exact position taken, it is implicit in how the research is carried out and how the results are communicated. This paper undertakes a qualitative analysis of the top 100 most cited papers from NeurIPS and ICML to uncover some of the most prominent values these papers espouse and how they shape the path forward. As we get a higher proliferation of AI in various aspects of our lives, critical scholars have raised concerns about the negative impacts of these systems on society. Yet, most technical papers published today pay little to no attention to the societal implications of their work. And this is despite emerging requirements like "Broader Impact Statements" that have become mandatory at several conferences. Through the manual analysis of 100 papers, this research surfaces trends that support this position and articulates that machine learning is not value-neutral.

MetroLab Network has partnered with Government Technology to bring its readers a segment called the MetroLab Innovation of the Month Series, which highlights impactful tech, data and innovation projects underway between cities and universities. If you'd like to learn more or contact the project leads, please contact MetroLab at info@metrolabnetwork.org for more information. In this month's installment of the Innovation of the Month series, we explore the work of Julia Stoyanovich, an assistant professor of Computer Science, Engineering, and Data Science at New York University, and Falaah Arif Khan from Data, Responsibly, who are creating comics designed to increase awareness of responsible data science. MetroLab's Ben Levine spoke with the two about the background and development of their project. Ben Levine: Can you tell us about the origin of the Data, Responsibly project and who has been involved in it?

A Tensor Processing Unit (TPU) is an Accelerator Application-Specific integrated Circuit (ASIC) developed by Google for Artificial Intelligence and Neural Network Machine Learning. With Machine Learning gaining its relevance and importance every day, the conventional microprocessors have known to be unable to effectively handle the computations be it training or neural network processing. The 1st Generation TPU is a hardware chip used at Google data center for faster computation. The 2nd generation TPU is now available in cloud and empowers businesses everywhere to access this accelerator technology to speed up their machine learning workloads using its high speed network. The 3rd generation TPU is twice as powerful as its previous generation and this result in an 8-fold increase in performance.

LONDON--(BUSINESS WIRE)--In a year when artificial intelligence and machine learning became a very hot topic in the media industry, the IBC Best Conference Paper Award goes to a team from BBC R&D which has investigated practical applications. The award is made to the technical paper which, according to the team of peer reviewers, delivers not just the most significant new research, but does so in an accessible way. The paper, 'AI in production: video analysis and machine learning for expanded live events coverage', will be presented at midday on Sunday 16 September as part of a new initiative at IBC2018 – 'Tech Talks'. 'Tech Talks' ensures that the highly respected technical papers remain an integral part of IBC and its conference, bringing the latest ideas to all delegates in a fresh and accessible form. Talking of the new innovation, Dr Nick Lodge, executive producer of technical sessions in the conference, said "Senior technologists and researchers who have been responsible for original and thought-provoking advances in media technology will talk about their own work, and audiences will have the rare opportunity to question these world experts. "The technologies that impact the media industry are broad," he added. "This year's'Tech Talks' will cover emerging areas like artificial intelligence, virtual and augmented reality, 5G and blockchain." In the award-winning paper a team of BBC researchers covering a wide range of skills, under project lead Mike Evans, discuss a project known as'Ed'. This prototype system is used to create near-live content with minimal crew. An example might be a set of three unmanned 4K cameras, from which'Ed' would produce a number of properly framed HD pictures, cutting between them as appropriate. "The point of the work is to allow coverage of more events, to reach places we otherwise could not reach," Mike Evans said. "With conventional production we cover only about six of the nearly 100 places music is performed at the Glastonbury Festival, for example, or just a tiny fraction of the 50,000 performances in 300 venues at the Edinburgh Fringe." "But with'Ed' we can reach many more of these and do so with production techniques which are much less intrusive for the event itself," he explained. "This technology will be suitable not just for major production companies like the BBC, but for a whole range of use cases, like minor sports which need to increase visibility, and even vloggers who want to improve their online presence." Dr Paul Entwistle, Chair of the IBC Technical Papers Committee which provides careful peer review of the many papers proposed for IBC, said "The detail in this paper is absolutely fascinating.

In this article, I report on the primary features of the IJCAI-07 program, including its theme, schedule, and organization. In particular, I discuss an effective and novel presentation format at IJCAI in which oral and poster papers were presented in the same sessions categorized by topic area. The theme of the conference was "AI and its benefits to society," with the aim of highlighting and raising our research interests for problems of direct relevance to society. It was our hope that the theme of the conference could also be discerned in the technical papers and that the work be used towards the realization of that goal. The theme was particularly evident in the invited talks.

On May 11, 1997, a computer showed that it could outclass a human in that most human of pursuits: playing a game. The human was World Chess Champion Garry Kasparov, and the computer was IBM's Deep Blue, which had begun life at Carnegie Mellon University as a system called ChipTest. One of Deep Blue's creators, Murray Campbell, talked to the IDG News Service about the other things computers have learned to do as well as, or better than, humans, and what that means for our future. What follows is an edited version of that conversation. IDGNS: Is it true that you and Deep Blue joined IBM at the same time?