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.

Instability and variability of Deep Reinforcement Learning (DRL) algorithms tend to adversely affect their performance. Averaged-DQN is a simple extension to the DQN algorithm, based on averaging previously learned Q-values estimates, which leads to a more stable training procedure and improved performance by reducing approximation error variance in the target values. To understand the effect of the algorithm, we examine the source of value function estimation errors and provide an analytical comparison within a simplified model. We further present experiments on the Arcade Learning Environment benchmark that demonstrate significantly improved stability and performance due to the proposed extension.

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.

The community is used by more than 800,000 data scientists. Google is set to acquire Kaggle, an Australia-based data science platform for running coding competition. The financial details about the acquisition have not been disclosed. Founded in 2010 by Anthony Goldbloom and Ben Hamner, Kaggle is a platform where data science and machine learning competitions are run. The platform is being used by more than 800,000 data scientists to explore, analyse and understand the latest updates in machine learning and data analytics.

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.

A new study shows that playing violent video games has no long term effects on a player's empathy. Despite what much of the mainstream media would have you believe, a new study has shown that playing violent videogames does not turn you into some kind of deranged psychopath. With more and more scientific studies examining the effects of video games on people, it's becoming clear that gaming is not only cognitively beneficial but has no long term detrimental effects in regards to a gamer's empathy. In a recent study published in Frontiers in Psychology, Dr. Gregor Szycik and his staff of the Hannover Medical School examined the brains of long term gamers and non-gamers with functional magnetic resonance imaging (fMRI). They saw that when both groups were exposed to emotionally provocative imagery that they reacted the same way, showing that playing games doesn't numb a gamer's empathy.

For the cognitively ambitious among us, there are always new debates to win, tests to ace, and feats of intellect to accomplish. And just as for the athletically gifted, the lure of performance-enhancing drugs is hard to resist. The question is, can drugs boost intellectual performance? New research shows that for tournament-level chess players, they can. In a series of experiments, researchers found that caffeine, the stimulant Ritalin and the stay-awake drug modafinil ever-so-slightly improved players' chances of winning a game against a computer program set to their skill level.

SANTA CLARA, CA--(Marketwired - Mar 8, 2017) - NVIDIA (NASDAQ: NVDA) with Microsoft today unveiled blueprints for a new hyperscale GPU accelerator to drive AI cloud computing. Providing hyperscale data centers with a fast, flexible path for AI, the new HGX-1 hyperscale GPU accelerator is an open-source design released in conjunction with Microsoft's Project Olympus. HGX-1 does for cloud-based AI workloads what ATX -- Advanced Technology eXtended -- did for PC motherboards when it was introduced more than two decades ago. It establishes an industry standard that can be rapidly and efficiently embraced to help meet surging market demand. The new architecture is designed to meet the exploding demand for AI computing in the cloud -- in fields such as autonomous driving, personalized healthcare, superhuman voice recognition, data and video analytics, and molecular simulations.

International Business Machines Corp. and Salesforce.com Inc. agreed to mingle their artificial-intelligence technologies in a bid to boost sales of the powerful data-analytics offerings. The companies Monday announced plans to offer integrated AI services that weave the broad humanlike conversation and learning capabilities of IBM's Watson with Salesforce's more sales-oriented Einstein technology. The new offerings, available in the second half of the year, are aimed at helping a wide variety of companies better target products and services at customers. IBM CEO: AI Will Be Man and Machine, Not Man vs. Machine. Ginni Rometty, CEO of IBM, said that she believes that believes artificial intelligence will help to create jobs and that clients will have a "symbiotic relationship" with AI.

NEW YORK The founder of former telecommunications company Global Crossing Inc, Gary Winnick, has acquired a majority stake in live streaming concert service Qello - the first step in creating a new company and technology platform to capitalize on the surge in online TV consumption, Winnick told Reuters.