Non-linear kernel methods can be approximated by fast linear ones using suitable explicit feature maps allowing their application to large scale problems. To this end, explicit feature maps of kernels for vectorial data have been extensively studied. As many real-world data is structured, various kernels for complex data like graphs have been proposed. Indeed, many of them directly compute feature maps. However, the kernel trick is employed when the number of features is very large or the individual vertices of graphs are annotated by real-valued attributes. Can we still compute explicit feature maps efficiently under these circumstances? Triggered by this question, we investigate how general convolution kernels are composed from base kernels and construct corresponding feature maps. We apply our results to widely used graph kernels and analyze for which kernels and graph properties computation by explicit feature maps is feasible and actually more efficient. In particular, we derive feature maps for random walk and subgraph matching kernels and apply them to real-world graphs with discrete labels. Thereby, our theoretical results are confirmed experimentally by observing a phase transition when comparing running time with respect to label diversity, walk lengths and subgraph size, respectively. Moreover, we derive approximative, explicit feature maps for state-of-the-art kernels supporting real-valued attributes including the GraphHopper and Graph Invariant kernels. In extensive experiments we show that our approaches often achieve a classification accuracy close to the exact methods based on the kernel trick, but require only a fraction of their running time.

UTOMATED segmentation of skin lesions in dermoscopy images is currently a challenging problem [1]. This paper proposes an innovative method to address this problem developed by the authors. It has been structured as follows. Firstly, in this introduction, on the one hand the segmentation problem is described and, on the other, the evaluation criteria used (image database, ground truths and metrics) are shown. Secondly, the system design is presented. Thirdly, the results and the discussion are shown. A. Problems with segmentation of skin lesions in dermoscopy images Automated segmentation of a skin lesion is a complex issue, as the possible casuistry that can appear in the images is very diverse. The main problems that can de found in the image which make segmentation difficult are as follows: 1. Presence of hair; 2. Other artifacts such as electronic letters, rulers, ink and color charts, etc.; 3. Dark rectangular or circular marks around it (a consequence of shadow); 4. Flashes; 5. Lighting problems: apart from the problem with dark marks and flashes that have already been mentioned, in some cases one part of the image turns out to be darker than another (a common cases is that the part of the skin beside the circular marks is often darker as it is less brightly lit, and some images also turn out to be darker than others; 6. As a result of the oil used to acquire many images, there may be distortion problems and bubbles; 7. Presence of blood vessels; 8. Presence of regression areas and blue-whitish veil -in many cases these structures have greater intensity than the skin surrounding the lesion; 9. Hypopigmentation areas which are confused with skin; 10.

The time to converge to the steady state of a finite Markov chain can be greatly reduced by a lifting operation, which creates a new Markov chain on an expanded state space. For a class of quadratic objectives, we show an analogous behavior where a distributed ADMM algorithm can be seen as a lifting of Gradient Descent algorithm. This provides a deep insight for its faster convergence rate under optimal parameter tuning. We conjecture that this gain is always present, as opposed to the lifting of a Markov chain which sometimes only provides a marginal speedup.

Variational inference using the reparameterization trick has enabled large-scale approximate Bayesian inference in complex probabilistic models, leveraging stochastic optimization to sidestep intractable expectations. The reparameterization trick is applicable when we can simulate a random variable by applying a differentiable deterministic function on an auxiliary random variable whose distribution is fixed. For many distributions of interest (such as the gamma or Dirichlet), simulation of random variables relies on acceptance-rejection sampling. The discontinuity introduced by the accept-reject step means that standard reparameterization tricks are not applicable. We propose a new method that lets us leverage reparameterization gradients even when variables are outputs of a acceptance-rejection sampling algorithm. Our approach enables reparameterization on a larger class of variational distributions. In several studies of real and synthetic data, we show that the variance of the estimator of the gradient is significantly lower than other state-of-the-art methods. This leads to faster convergence of stochastic gradient variational inference.

User preference integration is of great importance in multi-objective optimization, in particular in many objective optimization. Preferences have long been considered in traditional multicriteria decision making (MCDM) which is based on mathematical programming. Recently, it is integrated in multi-objective metaheuristics (MOMH), resulting in focus on preferred parts of the Pareto front instead of the whole Pareto front. The number of publications on preference-based multi-objective metaheuristics has increased rapidly over the past decades. There already exist various preference handling methods and MOMH methods, which have been combined in diverse ways. This article proposes to use the Web Ontology Language (OWL) to model and systematize the results developed in this field. A review of the existing work is provided, based on which an ontology is built and instantiated with state-of-the-art results. The OWL ontology is made public and open to future extension. Moreover, the usage of the ontology is exemplified for different use-cases, including querying for methods that match an engineering application, bibliometric analysis, checking existence of combinations of preference models and MOMH techniques, and discovering opportunities for new research and open research questions.

Anyone who's ever flown a drone of any sort will tell you that sooner or later, you're going to crash it. The question is how exactly you will go about doing this, and how much of the drone will be functional after it's happened. Most flying animals somewhat frustratingly don't have this problem: Birds and insects run into things occasionally (or all the time, for small bugs), and just shrug it off and keep on going, thanks to their biological design, which includes both stiffness and flexibility. Now roboticists at the EPFL, in Lausanne, Switzerland, are relying on these same qualities to design a highly resilient quadrotor that's impressively difficult to destroy. There are three primary strategies for designing drones with impact resistance.

The operator of the disaster-struck Fukushima No. 1 nuclear power plant said Thursday it will attempt to examine the inside of reactor 1 next Tuesday using a remote-controlled robot. The move follows a botched attempt by another self-propelled robot to take a look inside reactor 2, which had also sustained a meltdown after the March 11, 2011, Great East Japan Earthquake and tsunami. That robot became unable to move when it encountered debris and eventually could not be retrieved. These are the first attempts by Tokyo Electric Power Co. Holdings Inc. to examine the insides of the wrecked reactors since the nuclear disaster started. For the reactor 1 inspection, Tepco said the new robot will carry out a four-day probe inside the containment vessel.

A Volkswagen'Cedric' self-driving automobile is presented during the Volkswagen Group Shaping The Future / Create Innovation event ahead of the 87th Geneva International Motor Show on March 6, 2017 in Geneva, Switzerland. As many of the most innovative companies in the world race to bring autonomous vehicle solutions to market, a fierce debate has emerged in the industry about the best way to build those solutions. An AV must make countless tactical choices moment to moment to navigate through its environment, choices that are second nature to experienced human drivers: how fast to go, whether to stop at a traffic signal, whether to slow to let another vehicle merge, whether to change lanes to avoid a parked car. These are highly safety-critical decisions. They can mean the difference between life and death on the road, millions of times over, every day. Given the stakes, it is no surprise that the question as to which technological approach to apply here has taken on huge importance and inspired vigorous debate.

With the CIA's ability to'breach almost anything connected to the internet' made public, many citizens have begun questioning their devices. A clip has surfaced showing an anonymous woman asking Amazon's Alexa a series of questions - starting with'would you lie to me' and finishing with'Alexa, are you connected to the CIA?' The virtual assistant swiftly responded to the first question, but shutdown after it was interrogated about its connections with the US government agency. A video appeared on Reddit that has people questioning their virtual assistants. A woman is seen asking Amazon's Alexa a series of questions. ' 'I am not always right, but I would never intentionally lie to you or anyone else,' responded Alexa. Alexa answered, 'The Unites States Central Intelligence Agency, CIA'.

When John Forbes Nash, the Nobel Prize-winning mathematician, schizophrenic, and paranoid delusional, was asked how he could believe that space aliens had recruited him to save the world, he gave a simple response. "Because the ideas I had about supernatural beings came to me the same way that my mathematical ideas did. So I took them seriously." Nash is hardly the only so-called mad genius in history. Even ignoring those great creators who did not kill themselves in a fit of deep depression, it remains easy to list persons who endured well-documented psychopathology, including the composer Robert Schumann, the poet Emily Dickinson, and Nash.