The GANs are generative models whose random samples realistically reflect natural images. It also can generate samples with specific attributes by concatenating a condition vector into the input, yet research on this field is not well studied. We propose novel methods of conditioning generative adversarial networks (GANs) that achieve state-of-the-art results on MNIST and CIFAR-10. We mainly introduce two models: an information retrieving model that extracts conditional information from the samples, and a spatial bilinear pooling model that forms bilinear features derived from the spatial cross product of an image and a condition vector. These methods significantly enhance log-likelihood of test data under the conditional distributions compared to the methods of concatenation.

We investigate the problem of testing whether $d$ random variables, which may or may not be continuous, are jointly (or mutually) independent. Our method builds on ideas of the two variable Hilbert-Schmidt independence criterion (HSIC) but allows for an arbitrary number of variables. We embed the $d$-dimensional joint distribution and the product of the marginals into a reproducing kernel Hilbert space and define the $d$-variable Hilbert-Schmidt independence criterion (dHSIC) as the squared distance between the embeddings. In the population case, the value of dHSIC is zero if and only if the $d$ variables are jointly independent, as long as the kernel is characteristic. Based on an empirical estimate of dHSIC, we define three different non-parametric hypothesis tests: a permutation test, a bootstrap test and a test based on a Gamma approximation. We prove that the permutation test achieves the significance level and that the bootstrap test achieves pointwise asymptotic significance level as well as pointwise asymptotic consistency (i.e., it is able to detect any type of fixed dependence in the large sample limit). The Gamma approximation does not come with these guarantees; however, it is computationally very fast and for small $d$, it performs well in practice. Finally, we apply the test to a problem in causal discovery.

Attack graphs are a powerful tool for security risk assessment by analysing network vulnerabilities and the paths attackers can use to compromise network resources. The uncertainty about the attacker's behaviour makes Bayesian networks suitable to model attack graphs to perform static and dynamic analysis. Previous approaches have focused on the formalization of attack graphs into a Bayesian model rather than proposing mechanisms for their analysis. In this paper we propose to use efficient algorithms to make exact inference in Bayesian attack graphs, enabling the static and dynamic network risk assessments. To support the validity of our approach we have performed an extensive experimental evaluation on synthetic Bayesian attack graphs with different topologies, showing the computational advantages in terms of time and memory use of the proposed techniques when compared to existing approaches.

We consider the problem of maximizing a monotone submodular function under noise. There has been a great deal of work on optimization of submodular functions under various constraints, resulting in algorithms that provide desirable approximation guarantees. In many applications, however, we do not have access to the submodular function we aim to optimize, but rather to some erroneous or noisy version of it. This raises the question of whether provable guarantees are obtainable in presence of error and noise. We provide initial answers, by focusing on the question of maximizing a monotone submodular function under a cardinality constraint when given access to a noisy oracle of the function. We show that: - For a cardinality constraint $k \geq 2$, there is an approximation algorithm whose approximation ratio is arbitrarily close to $1-1/e$; - For $k=1$ there is an algorithm whose approximation ratio is arbitrarily close to $1/2$. No randomized algorithm can obtain an approximation ratio better than $1/2+o(1)$; -If the noise is adversarial, no non-trivial approximation guarantee can be obtained.

At 2:40 a.m., my phone woke me up. At least one of us was always on shift, and that night in September of 2010, I had volunteered to respond to automated text messages from our alert system. As a graduate student at the time, I (Jonah) had helped build the first quick-response alert software pipeline for two gravitational-wave observatories, called LIGO (Laser Interferometer Gravitational-wave Observatory) and Virgo. This system was designed to search for astrophysical signals in data as it arrived, to alert people who could check if a signal seemed valid, and share the message with astronomers around the world if needed. Every alert carried the possibility of a positive detection--humanity's first direct observation of waves traveling through the fabric of spacetime, predicted by Einstein in 1916. I got out of bed and made a sleepy-eyed walk to the small workstation we kept in our apartment. I didn't know it, but the alert was the beginning of a professional and emotional rollercoaster. I logged into our event database and started browsing plots. The plots showed an unusually loud signal.

Most patients have spent years bouncing from one doctor to another, building up huge dossiers of medical notes. Rare diseases typically take at least five years to correctly name, and sometimes up to 30, by which time it can be too late for effective treatment. "This is an inefficient, costly business," Dr Jurgen Schafer, who heads the German university's medical team, said at a media conference at IBM Zurich in October. "The computer is not going to replace the physician. But with this amount of data, it is completely clear that we don't need more physicians – we need more computer power."

The internet is about to become a vicious, chaotic battlefield, and Elon Musk says advanced A.I. could make the carnage even worse. According to a short exchange on Musk's Twitter today, the systems that keep the internet running are particularly vulnerable to simple, brute-force computing attacks -- the kind of cyberwarfare that artificial intelligence excel at. On October 21, an unknown group of hackers wiped out part of the internet in the United States and Europe with a massive Denial of Service attack. The hackers used a massive "botnet" -- linked computers able to perform coordinated functions -- of simple Internet of Things devices to relentlessly overload the servers at Dyn Systems, which provides DNS services to a huge number of websites, including Spotify, Twitter, Netflix, and Reddit. For the most part, cybersecurity authorities believe that a human being or group of people orchestrated and executed the attack, plugging in the botnet's targets and making sure their digital blows landed.

There are few places better positioned to become a "smart city" than Singapore. That's an easy statement to justify. Singapore is an island city-state just 30 miles across that has been governed by the same party for decades. Putting the implied democratic flaws to one side, the geography and political stability of Singapore have aided the city in preparing for the future. Two years ago, those preparations got a name: "Smart Nation," an ambitious program to push the city, its residents and its government into the digital age.

Each of us generates nearly a gigabyte of data per day. This huge volume of data contains an incredible amount of information that we are able to read and organize thanks to cognitive computing. Cognitive computing comes from a mashup of cognitive science -- the study of the human brain and how it functions -- and computer science. he goal of cognitive computing is to simulate human thought processes in a computerized model. Using self-learning algorithms that use data mining, pattern recognition and natural language processing, the computer can mimic the way the human brain works. The potential areas of application of cognitive computing are many.

Yesterday, Yandex announced that they launched something similar to the Google RankBrain - well, they didn't say that, I am. They launched what they call Palekh which is name of a Russian city, the flag of that city is of a firebird, which you can see in the image above. Why the firebird, well, it has a long tail and this algorithm aims at improving the quality of the results for long tail queries. Yandex told us that they handle about 100 million queries per day fall under the "long-tail" classification within their search engine. That is about 40% of all the queries performed on that search engine.