Advances in computational science offer a principled pipeline for predictive modeling of cardiovascular flows and aspire to provide a valuable tool for monitoring, diagnostics and surgical planning. Such models can be nowadays deployed on large patient-specific topologies of systemic arterial networks and return detailed predictions on flow patterns, wall shear stresses, and pulse wave propagation. However, their success heavily relies on tedious pre-processing and calibration procedures that typically induce a significant computational cost, thus hampering their clinical applicability. In this work we put forth a machine learning framework that enables the seamless synthesis of non-invasive in-vivo measurement techniques and computational flow dynamics models derived from first physical principles. We illustrate this new paradigm by showing how one-dimensional models of pulsatile flow can be used to constrain the output of deep neural networks such that their predictions satisfy the conservation of mass and momentum principles. Once trained on noisy and scattered clinical data of flow and wall displacement, these networks can return physically consistent predictions for velocity, pressure and wall displacement pulse wave propagation, all without the need to employ conventional simulators. A simple post-processing of these outputs can also provide a cheap and effective way for estimating Windkessel model parameters that are required for the calibration of traditional computational models. The effectiveness of the proposed techniques is demonstrated through a series of prototype benchmarks, as well as a realistic clinical case involving in-vivo measurements near the aorta/carotid bifurcation of a healthy human subject.

Today's high-stakes adversarial interactions feature attackers who constantly breach the ever-improving security measures. Deception mitigates the defender's loss by misleading the attacker to make suboptimal decisions. In order to formally reason about deception, we introduce the feature deception game (FDG), a domain-independent game-theoretic model and present a learning and planning framework. We make the following contributions. (1) We show that we can uniformly learn the adversary's preferences using data from a modest number of deception strategies. (2) We propose an approximation algorithm for finding the optimal deception strategy and show that the problem is NP-hard. (3) We perform extensive experiments to empirically validate our methods and results.

In statistical classification, machine learning, social and other sciences, a number of measures of association have been developed and used for assessing and comparing individual classifiers, raters, and their groups. Among the measures, we find the weighted kappa, extensively used by psychometricians, and the monotone and supremum correlation coefficients, prominently used by social scientists and statisticians. In this paper, we introduce, justify, and explore several new members of the class of functional correlation coefficients that naturally arise when comparing classifiers. We illustrate the performance of the coefficients by reanalyzing a number of confusion matrices that have appeared in the literature.

The Hello Work chain of public job-placement offices will facilitate job searches from home and improve its usability by employers by adopting more information technology. The changes are aimed at promoting smoother matches between job seekers and companies, said an official of the labor ministry, which runs the chain. Job seekers can currently only browse a limited amount of the network's job information on the internet and must visit a Hello Work office to get the full details. Starting in January, however, they will be able to access all of the information via personal computers and smartphones, after registering at one of the offices. By setting up personal pages on Hello Work's computer system, they will also be able to save their data on their search conditions and job offers.

Washington, DC – In the future, artificial intelligence could augment the background investigative work performed by humans, cutting the time it takes and providing a more realistic, in-depth and realistic profile of the individual, the technical director for research and development and technology transfer at the Defense Security Service's National Background Investigative Services said recently. Mark Nehmer spoke at the "Genius Machines: The New Age of Artificial Intelligence" event, hosted by Nextgov and Defense One in Arlington, Virginia. Millions of service members, federal employees and contractors receive background checks and are issued clearances on a periodic basis. There are several problems with the current system of background investigations, Nehmer said. The use of artificial intelligence, or AI, could significantly reduce the time it takes investigations and ease the strain on already-overworked personnel and reduce the backlog of cases, Nehmer said.

The rat sat still in the middle of her cage, moving only in response to my touch, and even then only as if in slow-motion. My subject, GRat66, was a few months old, and except for her long bare tail, fit neatly into my palm a few minutes earlier, when I injected a few drops of a potent opiate under her skin, near the belly. Now, her beady black eyes bulged as she faded into an opiate stupor. I was preparing to implant minuscule electrodes into the rat's brain. The opiate would serve as an analgesic before, during, and after the surgery. It was the fall of 2018 and I was hoping the results of the surgery would help answer some questions that had been tormenting me as I embarked on the sixth year of my Ph.D. in neuroscience. How do the parts of the brain controlling movement interact with those responsible for visual sensation? Why do neurons in the visual areas jolt to action when an animal moves, even in the dark?

The U.S Capitol is seen at sunrise. Imagine there is a group of people in Congress with more influence over whether laws are passed and rules are changed, than any official committee or subcommittee in the House and Senate. Now, imagine the members of this powerful group are not even members of Congress – in fact, they're not elected officials at all. Finally, imagine this group operates in secret, refuses to explain its decisions in detail to anyone, and has shown a consistent bias against free market principles. Unfortunately, you don't have to imagine this scenario.

As CIO of the World Health Organisation (WHO), Bernardo Mariano is responsible for the IT strategy at a United Nations agency that aims to attain the highest possible level of health for all citizens in its 194 member states. His key objectives are to deliver the WHO's global strategy on digital health and to provide guidelines to countries on how to navigate their digital transformation journeys. "We will do that by bringing together all the national regulatory agencies from the countries that have them and then start discussing issues such as AI regulation and sharing our experiences," Mariano tells CIO UK from the AI Everything conference in Dubai. "Some agencies have already implemented digital strategies and others have not, so we have to cross-fertilise ideas among the different regulatory agencies to make sure that they all are able to deal with and handle the added power but also maintain the quality of the healthcare existing ecosystem." To help its member states to achieve this, the WHO recently released its first recommendations for digital health tech use. They outline how nations can use tech to improve the health of citizens by providing proper training, securing data and coordinating systems with other digital health resources, while taking into account infrastructural limitations.

In this work, we study the robustness of a CNN+RNN based image captioning system being subjected to adversarial noises. We propose to fool an image captioning system to generate some targeted partial captions for an image polluted by adversarial noises, even the targeted captions are totally irrelevant to the image content. A partial caption indicates that the words at some locations in this caption are observed, while words at other locations are not restricted.It is the first work to study exact adversarial attacks of targeted partial captions. Due to the sequential dependencies among words in a caption, we formulate the generation of adversarial noises for targeted partial captions as a structured output learning problem with latent variables. Both the generalized expectation maximization algorithm and structural SVMs with latent variables are then adopted to optimize the problem. The proposed methods generate very successful at-tacks to three popular CNN+RNN based image captioning models. Furthermore, the proposed attack methods are used to understand the inner mechanism of image captioning systems, providing the guidance to further improve automatic image captioning systems towards human captioning.

Reinforcement learning (RL) is capable of managing wireless, energy-harvesting IoT nodes by solving the problem of autonomous management in non-stationary, resource-constrained settings. We show that the state-of-the-art policy-gradient approaches to RL are appropriate for the IoT domain and that they outperform previous approaches. Due to the ability to model continuous observation and action spaces, as well as improved function approximation capability, the new approaches are able to solve harder problems, permitting reward functions that are better aligned with the actual application goals. We show such a reward function and use policy-gradient approaches to learn capable policies, leading to behavior more appropriate for IoT nodes with less manual design effort, increasing the level of autonomy in IoT.