Once users have shared their data online, it is generally difficult for them to revoke access and ask for the data to be deleted. Machine learning (ML) exacerbates this problem because any model trained with said data may have memorized it, putting users at risk of a successful privacy attack exposing their information. Yet, having models unlearn is notoriously difficult. After a data point is removed from a training set, one often resorts to entirely retraining downstream models from scratch. We introduce SISA training, a framework that decreases the number of model parameters affected by an unlearning request and caches intermediate outputs of the training algorithm to limit the number of model updates that need to be computed to have these parameters unlearn. This framework reduces the computational overhead associated with unlearning, even in the worst-case setting where unlearning requests are made uniformly across the training set. In some cases, we may have a prior on the distribution of unlearning requests that will be issued by users. We may take this prior into account to partition and order data accordingly and further decrease overhead from unlearning. Our evaluation spans two datasets from different application domains, with corresponding motivations for unlearning. Under no distributional assumptions, we observe that SISA training improves unlearning for the Purchase dataset by 3.13x, and 1.658x for the SVHN dataset, over retraining from scratch. We also validate how knowledge of the unlearning distribution provides further improvements in retraining time by simulating a scenario where we model unlearning requests that come from users of a commercial product that is available in countries with varying sensitivity to privacy. Our work contributes to practical data governance in machine learning.

University of California, Davis Abstract--Despite the plethora of studies about security vulnerabilities and defenses of deep learning models, security aspects of deep learning methodologies, such as transfer learning, have been rarely studied. In this article, we highlight the security challenges and research opportunities of these methodologies, focusing on vulnerabilities and attacks unique to them. W ith the widespread adaptation of deep neural networks (DNN), their security challenges have received significant attention from both academia and industry, especially for mission critical applications, such as road sign detection for autonomous vehicles, face recognition in authentication systems, and fraud detection in financial systems. There are three major types of attacks on deep learning models, namely adversarial attacks, data poisoning, and exploratory attacks. Particularly, adversarial attacks, which aim to carefully craft inputs that cause the model to misclassify, has been extensively studied and many defence mechanisms have been proposed to alleviate them. These attacks are of paramount importance because they are effective, moderately simple to launch, and often transferable from one model to another. In literature, there are several survey and review papers on deep learning security and defence mechanisms. In this article, we focus on security of a much less explored area of machine learning - machine learning methodologies. Machine learning methodologies have been widely used to mitigate the restrictions and assumptions of a typical machine learning process. A typical DNN training process assumes large labeled dataset(s), access to high computational resources, non-private and centralized data, standard training and hyper-parameter tuning, and fixed task distribution over time. However, these assumptions are often difficult to realize in practice. Notwithstanding the proliferation of these machine learning methodologies, their security aspects have not been comprehensively analyzed, if ever studied. In this article, we focus on potential attacks, security vulnerabilities, and future directions specific to each learning methodology.

In the wake of the May 2018 mass shooting that resulted in 10 deaths at Santa Fe (Texas) High School, the Santa Fe Independent School District looked at all possible options to improve school safety within reasonable financial constraints. The district considered the idea of technology to enhance its approximately 750 cameras with facial recognition but did not immediately see a workable solution -- for reasons of cost, and concerns about shaky accuracy that could lead to false positives, says Kip Robins, director of technology for Santa Fe ISD, which has about 4,500 students. The district ultimately contracted with a company called AnyVision, which demonstrated its Better Tomorrow product, an artificial-intelligence-based application that plugs into an existing camera network and provides the ability to do surveillance based on a certain face, body or object. School districts or other end users can create a watch list to keep an eye out for potential pedophiles, for example, or someone known to be mentally unstable. The Santa Fe ISD's solution is part of a growing cadre of software offerings that use artificial intelligence to power through reams of data and notice certain predetermined visual information – whether it's someone's face, or a certain license plate, or simply human movement in a place and time where there shouldn't be any.

In this paper, we (1) argue that the international human rights framework provides the most promising set of standards for ensuring that AI systems are ethical in their design, development and deployment, and (2) sketch the basic contours of a comprehensive governance framework, which we call'human rights-centred design, deliberation and oversight', for ensuring that AI can be relied upon to operate in ways that will not violate human rights.

It's impressive that Generative models like Open AI's GPT-2 automatically create texts using limited input. But controlling the attributes (topics, context, sentiment) of these texts, and paragraphs need an extra layer of work that includes architectural modifications/specific data understanding, etc. This work is done by a team of professionals from Uber, Caltech, and the Hong Kong University of Science and Technology. They worked on the model and created the Plug and Play Language Model (PPLM), which takes one or two attributes classifier and combines it with a pre-trained language model.

Facial recognition technology has advanced swiftly in the last five years. As University of Texas at Dallas researchers try to determine how computers have gotten as good as people at the task, they are also shedding light on how the human brain sorts information. UT Dallas scientists have analyzed the performance of the latest echelon of facial recognition algorithms, revealing the surprising way these programs -- which are based on machine learning -- work. Their study, published online Nov. 12 in Nature Machine Intelligence, shows that these sophisticated computer programs -- called deep convolutional neural networks (DCNNs) -- figured out how to identify faces differently than the researchers expected. "For the last 30 years, people have presumed that computer-based visual systems get rid of all the image-specific information -- angle, lighting, expression and so on," said Dr. Alice O'Toole, senior author of the study and the Aage and Margareta Møller Professor in the School of Behavioral and Brain Sciences.

To learn more about the use of artificial intelligence at it may be applied to analyzing stocks and markets, I asked the CEO and originator of Ainstein AI about her work in this area. Suzanne Cook is a Wharton School graduate and a seven-time Institutional Investor All Star Analyst. Cook anticipates a new golden era of research - high frequency automated research - thanks to the trifecta of (1) cloud - cheaper and more accessible computing, (2) scale analytics - unifying vastly expanded data sets, and (3) autonomous pattern recognition via artificial intelligence." Here's how our conversation went: John Navin: When artificial intelligence experts talk about "natively intelligent portfolios," what exactly are they referring to? Suzanne Cook: Let's compare natively intelligent portfolios to the current portfolio offerings – not smart (analytics not built in), not in the cloud and not intuitive, as they lack visualizations.

If there's one platform that knows how to remain controversial, it's Facebook. In 2011, the company introduced a facial recognition feature which allowed users to tag others through suggestions displayed on photos. Moreover, the person in the photo was automatically notified if the uploader's privacy allowed to do so. This received a lot of criticism from privacy concerned users since it was giving away their identity without consent at times. A few weeks back, a court in Illinois even went as far to issue a ruling stating that users within the State could sue Facebook over its facial recognition technology.

The and Regional Deep Learning Market report gives a purposeful depiction of the area by the practice for research, amalgamation, and review of data taken from various sources. The market analysts have displayed the different sidelines of the area with a point on recognizing the top players (Amazon Web Services (AWS), Google, IBM, Intel, Micron Technology, Microsoft, Nvidia, Qualcomm, Samsung Electronics, Sensory Inc., Skymind, Xilinx, AMD, General Vision, Graphcore, Mellanox Technologies, Huawei Technologies, Fujitsu, Baidu, Mythic, Adapteva, Inc., Koniku) of the industry. The and Regional Deep Learning market report correspondingly joins a predefined business market from a SWOT investigation of the real players. Thus, the data summarized out is, no matter how you look at it is, reliable and the result of expansive research. This report mulls over and Regional Deep Learning showcase on the classification, for instance, application, concords, innovations, income, improvement rate, import, and others (Automotive, Home & Building Automation, Food & Beverages) in the estimated time from 2019–2025 on a global stage.

Artificial Intelligence (AI) -- the ability of machines to make decisions that normally require human expertise -- already is changing our world in countless ways, from self-driving cars to facial-recognition technology. But the best -- and maybe the worst -- is yet to come. AI is being used increasingly in health care, including the possibility of a radiology tool that might eliminate the need for tissue samples. Knowing that, the people leading a new project called Ethical-AI for the Center for Practical Bioethics (CPB) are trying to make sure that AI health care tools will be created and used in ethical ways. The ethical questions the project is raising should have been considered in a systematic way years ago, of course.