Researchers have revealed the first ever liquid magnet that can stay magnetic even when changing its shape -- an attractive prospect for developing fluid robots. The liquid is made of nano-scale particles of metal floating in solution -- which normally would only behave as a magnet when in the presence of a magnetic field. But by using a special oil-polymer mixture, the team succeeded in jamming the particles so close together at the surface of the liquid that they can stay magnetic. The pioneering discovery changes our understand of magnetic materials and could find manifold practical applications in the future. Researchers have revealed the first ever liquid magnet that can stay magnetic even when changing its shape -- an attractive prospect for developing fluid robots. University of Massachusetts material scientist Thomas Russell and his colleagues spent seven years developing a simple method to transform so-called'paramagnetic ferrofluids' -- plain metal particles floating in a liquid -- into permanent magnets.

Iran's Revolutionary Guard claims the vessel was caught trying to smuggle Iranian oil to foreign ships; Trey Yingst reports. Iran on Friday denied President Trump's claim that a U.S. warship destroyed an Iranian drone near the Persian Gulf after it threatened the ship -- an incident that further escalated tensions between the countries. Trump said Thursday that the USS Boxer – which is among several U.S. Navy ships in the area – took defensive action after an Iranian drone came within 1,000 yards of the warship and ignored multiple calls to stand down. Trump blamed Iran for a "provocative and hostile" action and said the U.S. responded in self-defense. But Iran's foreign minister, Mohammad Javad Zarif, told reporters as he arrived for a meeting at the United Nations that "we have no information about losing a drone today."

WASHINGTON - A U.S. warship on Thursday destroyed an Iranian drone in the Strait of Hormuz after it threatened the ship, President Donald Trump said. The incident marked a new escalation of tensions between the countries less than one month after Iran downed an American drone in the same waterway and Trump came close to retaliating with a military strike. In remarks at the White House, Trump blamed Iran for a "provocative and hostile" action and said the U.S. responded in self-defense. He said the Navy's USS Boxer, an amphibious assault ship, took defensive action after the Iranian aircraft closed to within 1,000 yards of the ship and ignored multiple calls to stand down. "The United States reserves the right to defend our personnel, facilities and interests and calls upon all nations to condemn Iran's attempts to disrupt freedom of navigation and global commerce," Trump said.

Intelligent algorithms are recently used in the optimization process in chemical engineering and application of multiphase flows such as bubbling flow. This overview of modeling can be a great replacement with complex numerical methods or very time-consuming and disruptive measurement experimental process. In this study, we develop the adaptive network-based fuzzy inference system (ANFIS) method for mapping inputs and outputs together and understand the behavior of the fluid flow from other output parameters of the bubble column reactor. Neural cells can fully learn the process in their memory and after the training stage, the fuzzy structure predicts the multiphase flow data. Four inputs such as x coordinate, y coordinate, z coordinate, and air superficial velocity and one output such as pressure gradient are considered in the learning process of the ANFIS method. During the learning process, the different number of the membership function, type of membership functions and the number of inputs are examined to achieve the intelligent algorithm with high accuracy. The results show that as the number of inputs increases the accuracy of the ANFIS method rises up to R^2>0.99 almost for all cases, while the increment in the number of rules has a effect on the intelligence of artificial algorithm. This finding shows that the density of neural objects or higher input parameters enables the moded for better understanding. We also proposed a new evaluation of data in the bubble column reactor by mapping inputs and outputs and shuffle all parameters together to understand the behaviour of the multiphase flow as a function of either inputs or outputs. This new process of mapping inputs and outputs data provides a framework to fully understand the flow in the fluid domain in a short time of fuzzy structure calculation.

We provide recovery guarantees for compressible signals that have been corrupted with noise and extend the framework introduced in [1] to defend neural networks against $\ell_0$-norm and $\ell_2$-norm attacks. Concretely, for a signal that is approximately sparse in some transform domain and has been perturbed with noise, we provide guarantees for accurately recovering the signal in the transform domain. We can then use the recovered signal to reconstruct the signal in its original domain while largely removing the noise. Our results are general as they can be directly applied to most unitary transforms used in practice and hold for both $\ell_0$-norm bounded noise and $\ell_2$-norm bounded noise. In the case of $\ell_0$-norm bounded noise, we prove recovery guarantees for Iterative Hard Thresholding (IHT) and Basis Pursuit (BP). For the case of $\ell_2$-norm bounded noise, we provide recovery guarantees for BP. These guarantees theoretically bolster the defense framework introduced in [1] for defending neural networks against adversarial inputs. Finally, we experimentally demonstrate this defense framework using both IHT and BP against the One Pixel Attack [21], Carlini-Wagner $\ell_0$ and $\ell_2$ attacks [3], Jacobian Saliency Based attack [18], and the DeepFool attack [17] on CIFAR-10 [12], MNIST [13], and Fashion-MNIST [27] datasets. This expands beyond the experimental demonstrations of [1].

This paper proposes novel algorithms for speaker embedding using subjective inter-speaker similarity based on deep neural networks (DNNs). Although conventional DNN-based speaker embedding such as a $d$-vector can be applied to multi-speaker modeling in speech synthesis, it does not correlate with the subjective inter-speaker similarity and is not necessarily appropriate speaker representation for open speakers whose speech utterances are not included in the training data. We propose two training algorithms for DNN-based speaker embedding model using an inter-speaker similarity matrix obtained by large-scale subjective scoring. One is based on similarity vector embedding and trains the model to predict a vector of the similarity matrix as speaker representation. The other is based on similarity matrix embedding and trains the model to minimize the squared Frobenius norm between the similarity matrix and the Gram matrix of $d$-vectors, i.e., the inter-speaker similarity derived from the $d$-vectors. We crowdsourced the inter-speaker similarity scores of 153 Japanese female speakers, and the experimental results demonstrate that our algorithms learn speaker embedding that is highly correlated with the subjective similarity. We also apply the proposed speaker embedding to multi-speaker modeling in DNN-based speech synthesis and reveal that the proposed similarity vector embedding improves synthetic speech quality for open speakers whose speech utterances are unseen during the training.

It is typical for a machine learning system to have numerous hyperparameters that affect its learning rate and prediction quality. Finding a good combination of the hyperparameters is, however, a challenging job. This is mainly because evaluation of each combination is extremely expensive computationally; indeed, training a machine learning system on real data with just a single combination of hyperparameters usually takes hours or even days. In this paper, we address this challenge by trying to predict the performance of the machine learning system with a given combination of hyperparameters without completing the expensive learning process. Instead, we terminate the training process at an early stage, collect the model performance data and use it to predict which of the combinations of hyperparameters is most promising. Our preliminary experiments show that such a prediction improves the performance of the commonly used random search approach.

We provide a novel convergence analysis of two popular sampling algorithms, Weighted Kernel Herding and Sequential Bayesian Quadrature, that are used to approximate the expectation of a function under a distribution. Existing theoretical analysis was insufficient to explain the empirical successes of these algorithms. We improve upon existing convergence rates to show that, under mild assumptions, these algorithms converge linearly. To this end, we also suggest a simplifying assumption that is true for most cases in finite dimensions, and that acts as a sufficient condition for linear convergence to hold in the much harder case of infinite dimensions. When this condition is not satisfied, we provide a weaker convergence guarantee. Our analysis also yields a new distributed algorithm for large-scale computation that we prove converges linearly under the same assumptions. Finally, we provide an empirical evaluation to test the proposed algorithm for a real world application.

There is some theoretical evidence that deep neural networks with multiple hidden layers have a potential for more efficient representation of multidimensional mappings than shallow networks with a single hidden layer. The question is whether it is possible to exploit this theoretical advantage for finding such representations with help of numerical training methods. Tests using prototypical problems with a known mean square minimum did not confirm this hypothesis. Minima found with the help of deep networks have always been worse than those found using shallow networks. This does not directly contradict the theoretical findings---it is possible that the superior representational capacity of deep networks is genuine while finding the mean square minimum of such deep networks is a substantially harder problem than with shallow ones.

Predicting the remaining useful life of machinery, infrastructure, or other equipment can facilitate preemptive maintenance decisions, whereby a failure is prevented through timely repair or replacement. This allows for a better decision support by considering the anticipated time-to-failure and thus promises to reduce costs. Here a common baseline may be derived by fitting a probability density function to past lifetimes and then utilizing the (conditional) expected remaining useful life as a prognostic. This approach finds widespread use in practice because of its high explanatory power. A more accurate alternative is promised by machine learning, where forecasts incorporate deterioration processes and environmental variables through sensor data. However, machine learning largely functions as a black-box method and its forecasts thus forfeit most of the desired interpretability. As our primary contribution, we propose a structured-effect neural network for predicting the remaining useful life which combines the favorable properties of both approaches: its key innovation is that it offers both a high accountability and the flexibility of deep learning. The parameters are estimated via variational Bayesian inferences. The different approaches are compared based on the actual time-to-failure for aircraft engines. This demonstrates the performance and superior interpretability of our method, while we finally discuss implications for decision support.