When I was growing up, all my friends thought making video games would be just about the coolest thing you could do as a career. It would be a way to stay connected to the industry into adulthood, as surely, there was no other way to do so, right? Between YouTube and Twitch, actually playing video games for a living can be a solid source of income, a full-time job, or even a winning lottery ticket. YouTubers and Twitch streamers are now some of the biggest, most well-known icons in the video game industry, and many aspire to be just like them. I wanted to talk to someone in the community that wasn't making PewDiePie mountains of cash, but is still well-known and considering doing this full time.

Real-time agent-centered heuristic search is a well-studied problem where an agent that can only reason locally about the world must travel to a goal location using bounded computation and memory at each step. Many algorithms have been proposed for this problem and theoretical results have also been derived for the worst-case performance with simple examples demonstrating worst-case performance in practice. Lower bounds, however, have not been widely studied. In this paper we study best-case performance more generally and derive theoretical lower bounds for reaching the goal using LRTA*, a canonical example of a real-time agent-centered heuristic search algorithm. The results show that, given some reasonable restrictions on the state space and the heuristic function, the number of steps an LRTA*-like algorithm requires to reach the goal will grow asymptotically faster than the state space, resulting in ``scrubbing'' where the agent repeatedly visits the same state. We then show that while the asymptotic analysis does not hold for more complex real-time search algorithms, experimental results suggest that it is still descriptive of practical performance.

Non-stationary continuous time Bayesian networks are introduced. They allow the parents set of each node to change over continuous time. Three settings are developed for learning non-stationary continuous time Bayesian networks from data: known transition times, known number of epochs and unknown number of epochs. A score function for each setting is derived and the corresponding learning algorithm is developed. A set of numerical experiments on synthetic data is used to compare the effectiveness of non-stationary continuous time Bayesian networks to that of non-stationary dynamic Bayesian networks. Furthermore, the performance achieved by non-stationary continuous time Bayesian networks is compared to that achieved by state-of-the-art algorithms on four real-world datasets, namely drosophila, saccharomyces cerevisiae, songbird and macroeconomics.

Philosophers of science are not known for agreeing with each other--contrariness is part of the job description. But for thousands of years, from Aristotle to Thomas Kuhn, those who study what science is have roughly categorized themselves into two basic camps: "realists" and "anti-realists." In philosophical terms, "anti-realists" or "empiricists" understand science as investigating the properties of observable objects via experiments. Empirical theories are constrained by the experimental results. "Realists," on the other hand, speculate more freely about the possible shape of the unobservable world, often designing mathematical explanations that cannot (yet) be tested. Isaac Newton was a realist, as are string theorists. Most scientists do not lose sleep worrying about philosophical divides. But maybe they should; Albert Einstein certainly did, as did Niels Bohr, and Erwin Schrödinger.

This paper extends the standard chaining technique (e.g., Pollard, 1990; Dudley, 1999; Györfi et al., 2002; Boucheron et al., 2012) to prove high-probability excess risk upper bounds for empirical risk minimization (ERM) for random design settings even if the magnitude of the noise and the estimates is unbounded. Our result (Theorem 1) covers bounded settings (Bartlett et al., 2005; Koltchinskii, 2011), extends to sub-Gaussian or even subexponential noise(van de Geer, 2000; Györfi and Wegkamp, 2008), and handles hypothesis classes with unbounded magnitude (Lecué and Mendelson, 2013; Mendelson, 2014; Liang et al., 2015). Furthermore, it applies to many loss functions besides the squared loss, and does not need additional statistical assumptions such as the bounded kurtosis of the transformed covariates over the hypothesis class, which prevent the latest developments to provide tight excess risk bounds for many sub-Gaussian cases (Section 1.2). To demonstrate the effectiveness of our method for such unbounded settings, we use our general excess risk bound (Theorem 1) to provide a detailed analysis for linear least squares estimators using quadratic slope constraint and penalty with sub-Gaussian noise and domain for the random design, nonrealizable setting(Section 3). Our result for the slope constrained case extends Theorem A of Lecué and Mendelson (2013) and nearly proves the conjecture of Shamir (2015), while our treatment for the penalized case (ridge regression) is comparable to the work of Hsu et al. (2014). The rest of this section introduces our notation through the formal definition of the regression problem and ERM estimators (Section 1.1), and discusses the limitations of 1 current excess risk upper bounds in the literature (Section 1.2). Then, we provide our main result in Section 2 to upper bound the excess risk of ERM estimators, and discuss its properties for various settings including many loss functions besides the squared loss. Next, Section 3 provides a detailed analysis for linear least squares estimators including the slope constrained case (Section 3.1) and ridge regression (Section 3.2). Finally, Section 4 proves our main result (Theorem 1).

Siamak Ravanbakhsh, Barnab as P oczos, Jeff Schneider 1 and Dale Schuurmans, Russell Greiner 2 1 Carnegie Mellon University, 5000 Forbes Ave, Pittsburgh, PA 15213 2 University of Alberta, Edmonton, AB T6G 2E8, Canada Abstract We propose a Laplace approximation that creates a stochastic unit from any smooth monotonic activation function, using only Gaussian noise. This paper investigates the application of this stochastic approximation in training a family of Restricted Boltzmann Machines (RBM) that are closely linked to Bregman divergences. This family, that we call exponential family RBM (Exp-RBM), is a subset of the exponential family Harmoniums that expresses family members through a choice of smooth monotonic non-linearity for each neuron. Using contrastive divergence along with our Gaussian approximation, we show that Exp-RBM can learn useful representations using novel stochastic units. 1 Introduction Deep neural networks (LeCun et al., 2015; Bengio, 2009) have produced some of the best results in complex pattern recognition tasks where the training data is abundant. Here, we are interested in deep learning for generative modeling. Recent years has witnessed a surge of interest in directed generative models that are trained using (stochastic) back-propagation ( e.g., Kingma and Welling, 2013; Rezende et al., 2014; Goodfellow et al., 2014). These models are distinct from deep energy-based models - including deep Boltzmann machine (Hinton et al., 2006) and (convolutional) deep belief networkAppearing in Proceedings of the 19th International Conference on Artificial Intelligence and Statistics (AISTATS) 2016, Cadiz, Spain. Although, due to their use of Gaussian noise, the stochastic units that we introduce in this paper can be potentially used with stochastic back-propagation, this paper is limited to applications in RBM.

Deloitte said that around 3.3 million jobs could be classified as business services roles, and that of those, there was a "high chance" that 800,000 to one million jobs would no longer be performed by humans over the period. Simon Barnes, a Deloitte partner, said that the sector's workforce would "fundamentally change over the next 10 to 20 years". Humans are likely to be liberated from "repetitive and highly structured" roles, while new higher-skilled positions are expected to be created to replace them. Mark Carney, the Bank of England Governor, said last month that many of the jobs and industries we are now familiar with "will be gone tomorrow". The rising speed of technological change threatens to make it difficult to choose a career, and for young people to plan their lives, he said.

The increasing interest in complex networks research has been a consequence of several intrinsic features of this area, such as the generality of the approach to represent and model virtually any discrete system, and the incorporation of concepts and methods deriving from many areas, from statistical physics to sociology, which are often used in an independent way. Yet, for this same reason, it would be desirable to integrate these various aspects into a more coherent and organic framework, which would imply in several benefits normally allowed by the systematization in science, including the identification of new types of problems and the cross-fertilization between fields. More specifically, the identification of the main areas to which the concepts frequently used in complex networks can be applied paves the way to adopting and applying a larger set of concepts and methods deriving from those respective areas. Among the several areas that have been used in complex networks research, pattern recognition, optimization, linear algebra, and time series analysis seem to play a more basic and recurrent role. In the present manuscript, we propose a systematic way to integrate the concepts from these diverse areas regarding complex networks research. In order to do so, we start by grouping the multidisciplinary concepts into three main groups, namely features, similarity, and network connectivity. Then we show that several of the analysis and modeling approaches to complex networks can be thought as a composition of maps between these three groups, with emphasis on nine main types of mappings, which are presented and illustrated. Such a systematization of principles and approaches also provides an opportunity to review some of the most closely related works in the literature, which is also developed in this article.

Dr. Richard Sutton presents "The Future of Artificial Intelligence" in the Technology and Future of Medicine course LABMP 590 http://www.singularitycourse.com This video has the greatest potential to save the world, and improve everyone's preparation for the future and improve the actual future itself, of any videos we have produced to date. If you do not want to watch the whole thing from the beginning, watch from 00:27:24 The Enslavement Problem. This is from the September 10th, 2015 lecture in eHub at the University of Alberta in Edmonton, Canada. Dr. Kim Solez gave a poetry reading on subjects related to this lecture at the Strathearn Art Walk on September 12, 2015, with Dr. Sutton in the audience and commenting afterward see https://www.youtube.com/watch?v jTVO7... .

This isn't the first piece of music composed by a computer. But it's a moment that might one day be called "watershed." It might also be called "the beginning of the end." Google unveiled its Magenta program at Moogfest. The program, part of the deep learning Google Brain project, will use Google's open-source artificial intelligence platform TensorFlow to research machine learning in artistic creation, i.e. have a machine write a song.