We develop and study new adversarial perturbations that enable an attacker to gain control over decisions in generic Artificial Intelligence (AI) systems including deep learning neural networks. In contrast to adversarial data modification, the attack mechanism we consider here involves alterations to the AI system itself. Such a stealth attack could be conducted by a mischievous, corrupt or disgruntled member of a software development team. It could also be made by those wishing to exploit a "democratization of AI" agenda, where network architectures and trained parameter sets are shared publicly. Building on work by [Tyukin et al., International Joint Conference on Neural Networks, 2020], we develop a range of new implementable attack strategies with accompanying analysis, showing that with high probability a stealth attack can be made transparent, in the sense that system performance is unchanged on a fixed validation set which is unknown to the attacker, while evoking any desired output on a trigger input of interest. The attacker only needs to have estimates of the size of the validation set and the spread of the AI's relevant latent space. In the case of deep learning neural networks, we show that a one neuron attack is possible - a modification to the weights and bias associated with a single neuron - revealing a vulnerability arising from over-parameterization. We illustrate these concepts in a realistic setting. Guided by the theory and computational results, we also propose strategies to guard against stealth attacks.

After decades of uninterrupted progress and growth, information technology has so evolved that it can be said we are entering the age of autonomous machines, but there exist many roadblocks in the way of making this a reality. In this article, we make a preliminary attempt at recognizing and categorizing the technical and non-technical challenges of autonomous machines; for each of the ten areas we have identified, we review current status, roadblocks, and potential research directions. It is hoped that this will help the community define clear, effective, and more formal development goalposts for the future.

The common practice of quality monitoring in industry relies on manual inspection well-known to be slow, error-prone and operator-dependent. This issue raises strong demand for automated real-time quality monitoring developed from data-driven approaches thus alleviating from operator dependence and adapting to various process uncertainties. Nonetheless, current approaches do not take into account the streaming nature of sensory information while relying heavily on hand-crafted features making them application-specific. This paper proposes the online quality monitoring methodology developed from recently developed deep learning algorithms for data streams, Neural Networks with Dynamically Evolved Capacity (NADINE), namely NADINE++. It features the integration of 1-D and 2-D convolutional layers to extract natural features of time-series and visual data streams captured from sensors and cameras of the injection molding machines from our own project. Real-time experiments have been conducted where the online quality monitoring task is simulated on the fly under the prequential test-then-train fashion - the prominent data stream evaluation protocol. Comparison with the state-of-the-art techniques clearly exhibits the advantage of NADINE++ with 4.68\% improvement on average for the quality monitoring task in streaming environments. To support the reproducible research initiative, codes, results of NADINE++ along with supplementary materials and injection molding dataset are made available in \url{https://github.com/ContinualAL/NADINE-IJCNN2021}.

We present a novel multi-view training framework and CNN architecture for combining information from multiple overlapping satellite images and noisy training labels derived from OpenStreetMap (OSM) to semantically label buildings and roads across large geographic regions (100 km$^2$). Our approach to multi-view semantic segmentation yields a 4-7% improvement in the per-class IoU scores compared to the traditional approaches that use the views independently of one another. A unique (and, perhaps, surprising) property of our system is that modifications that are added to the tail-end of the CNN for learning from the multi-view data can be discarded at the time of inference with a relatively small penalty in the overall performance. This implies that the benefits of training using multiple views are absorbed by all the layers of the network. Additionally, our approach only adds a small overhead in terms of the GPU-memory consumption even when training with as many as 32 views per scene. The system we present is end-to-end automated, which facilitates comparing the classifiers trained directly on true orthophotos vis-a-vis first training them on the off-nadir images and subsequently translating the predicted labels to geographical coordinates. With no human supervision, our IoU scores for the buildings and roads classes are 0.8 and 0.64 respectively which are better than state-of-the-art approaches that use OSM labels and that are not completely automated.

Hyundai Motor Company, headquartered in Seoul, South Korea, is one of the largest car manufacturers in the world. They have been heavily investing human and material resources in the race to develop self-driving cars, also known as autonomous vehicles. One of the algorithms often used in autonomous driving is semantic segmentation, which is a task to annotate every pixel of an image with an object class. These classes could be road, person, car, building, vegetation, sky, and so on. In a typical development cycle, the team at Hyundai Motor Company tests the accuracy periodically, and gathers additional images to correct for the insufficient predictive performance in specific situations.

AWS customers are increasingly using computer vision (CV) models for improved efficiency and an enhanced user experience. For example, a live broadcast of sports can be processed in real time to detect specific events automatically and provide additional insights to viewers at low latency. Inventory inspection at large warehouses capture and process millions of images across their network to identify misplaced inventory. CV models can be built with multiple deep learning frameworks like TensorFlow, PyTorch, and Apache MXNet. These models typically have a large input payload of images or videos of varying size.

The researchers offered concrete evidence of that risk by conducting a case study of a deep learning model that analyzes chest X-rays for signs of collapsed lungs. The system was trained and tested on patient data from Stanford Health Center, but Zou and his colleagues tested it against patient data from two other sites -- the National Institute of Health in Bethesda, Md., and Beth Israel Deaconess Medical Center in Boston. Sure enough, the algorithms were almost 10 percent less accurate at the other sites. In Boston, moreover, they found that their accuracy was higher for white patients than for Black patients. AI systems have been famously vulnerable to built-in racial and gender biases, Zou notes.

Nvidia has unveiled an AI model that converts a single 2D image of a person into a "talking head" video. Known as Vid2Vid Cameo, the deep learning model is designed to improve the experience of videoconferencing. If you're running late for a call, you could roll out of bed in your pajamas and disheveled hair, upload a photo of you dressed to impress, and the AI will map your facial movements to the reference image -- leaving the other attendees unaware of the chaos behind the camera. That could be a boon for the chronically unkempt, but you should probably test the technique before you turn up in your birthday suit. The system can also adjust your talking head's viewpoint to show you looking straight at the screen, when secretly your eyes are fixed on a TV in the background.

This article is part of Demystifying AI, a series of posts that (try to) disambiguate the jargon and myths surrounding AI. Today, deep learning might seem like a manifestation of the saying by British science fiction writer Arthur C. Clarke: "Any sufficiently advanced technology is indistinguishable from magic." In the past years, deep learning has proven to be capable of creating realistic images of non-existing people, recognize faces and voice commands, synthesize voice that sounds almost natural, and (pretend to) understand natural language. The remarkable feats of deep learning make it seem magical and out of reach. Yet, at heart, any deep learning model is just a combination of simple mathematical components.

Thanks to the exponential growth of malware, traditional heuristics-based detection regimes have been overwhelmed, leaving computers at risk. Machine learning approaches can help, but the bottleneck presented by the feature engineering step is a potential dealbreaker. The best path forward at this point is deep learning, says the CEO of Deep Instinct, which claims to have taken an early lead in the emerging field. Ten years ago, the cybersecurity industry faced a dilemma. The volume of malware was exploding, with tens of thousands of new types discovered every day.