Microsoft last week apologized for its Tay chatbot's bad behavior. It took the machine learning system offline, only 24 hours into its short life, after Twitter trolls got it to deny the Holocaust and elicit pro-Nazi and anti-feminist remarks. "We are deeply sorry for the unintended offensive and hurtful tweets from Tay, which do not represent who we are or what we stand for, nor how we designed Tay," said Peter Lee, corporate vice president at Microsoft Research. The company launched Tay on Twitter with the goal of learning about and improving the artificial intelligence by having it interact with 18- to 24-year-old U.S. Web users. Microsoft says an e-gang forced it to take Tay down.

Artificial Intelligence (AI) has been much in the news with the defeat of Go champion Lee Sedol. In the past, some including Stephen Hawking and numerous Hollywood films have warned about the inexorable march of a potentially hostile Artificial Intelligence. Others fear the disruption to traditional jobs of these new technologies. And if you don't work in the tech industry, it's easy to get the impression that AI systems are just ticking things off the list one by one until we humans are all redundant. But human vs. machine competitions that capture our imaginations represent only a small slice of the amazing AI research that's occurring worldwide.

Last year, a curious nonfiction book became a Times best-seller: a dense meditation on artificial intelligence by the philosopher Nick Bostrom, who holds an appointment at Oxford. Titled "Superintelligence: Paths, Dangers, Strategies," it argues that true artificial intelligence, if it is realized, might pose a danger that exceeds every previous threat from technology--even nuclear weapons--and that if its development is not managed carefully humanity risks engineering its own extinction. Central to this concern is the prospect of an "intelligence explosion," a speculative event in which an A.I. gains the ability to improve itself, and in short order exceeds the intellectual potential of the human brain by many orders of magnitude. Such a system would effectively be a new kind of life, and Bostrom's fears, in their simplest form, are evolutionary: that humanity will unexpectedly become outmatched by a smarter competitor. He sometimes notes, as a point of comparison, the trajectories ...

With retro games and Go well-conquered, where is an artificial intelligence like Google DeepMind meant to turn next? Before you get too excited (or maybe insanely depressed as you imagine a toaster holding aloft the Magic World Championship trophy on its ejection lever), there are no plans to set the AI loose on playing these popular card games. For now, the folks over at Oxford University are happy enough for DeepMind to analyse card data and transform it into code. Essentially, the task it is being set is one of translating the data from human to machine speak and while the cards have their own game "language" and structure, they can certainly throw some curveballs. Many language generation tasks require the production of text conditioned on both structured and unstructured inputs. We present a novel neural network architecture which generates an output sequence conditioned on an arbitrary number of input functions.

Brain decoding is a popular multivariate approach for hypothesis testing in neuroimaging. It is well known that the brain maps derived from weights of linear classifiers are hard to interpret because of high correlations between predictors, low signal to noise ratios, and the high dimensionality of neuroimaging data. Therefore, improving the interpretability of brain decoding approaches is of primary interest in many neuroimaging studies. Despite extensive studies of this type, at present, there is no formal definition for interpretability of multivariate brain maps. As a consequence, there is no quantitative measure for evaluating the interpretability of different brain decoding methods. In this paper, first, we present a theoretical definition of interpretability in brain decoding; we show that the interpretability of multivariate brain maps can be decomposed into their reproducibility and representativeness. Second, as an application of the proposed theoretical definition, we formalize a heuristic method for approximating the interpretability of multivariate brain maps in a binary magnetoencephalography (MEG) decoding scenario. Third, we propose to combine the approximated interpretability and the performance of the brain decoding model into a new multi-objective criterion for model selection. Our results for the MEG data show that optimizing the hyper-parameters of the regularized linear classifier based on the proposed criterion results in more informative multivariate brain maps. More importantly, the presented definition provides the theoretical background for quantitative evaluation of interpretability, and hence, facilitates the development of more effective brain decoding algorithms in the future.

Asymptotic factorizations for the small-ball probability (SmBP) of a Hilbert valued random element $X$ are rigorously established and discussed. In particular, given the first $d$ principal components (PCs) and as the radius $\varepsilon$ of the ball tends to zero, the SmBP is asymptotically proportional to (a) the joint density of the first $d$ PCs, (b) the volume of the $d$-dimensional ball with radius $\varepsilon$, and (c) a correction factor weighting the use of a truncated version of the process expansion. Moreover, under suitable assumptions on the spectrum of the covariance operator of $X$ and as $d$ diverges to infinity when $\varepsilon$ vanishes, some simplifications occur. In particular, the SmBP factorizes asymptotically as the product of the joint density of the first $d$ PCs and a pure volume parameter. All the provided factorizations allow to define a surrogate intensity of the SmBP that, in some cases, leads to a genuine intensity. To operationalize the stated results, a non-parametric estimator for the surrogate intensity is introduced and it is proved that the use of estimated PCs, instead of the true ones, does not affect the rate of convergence. Finally, as an illustration, simulations in controlled frameworks are provided.

In statistical genetics an important task involves building predictive models for the genotype-phenotype relationships and thus attribute a proportion of the total phenotypic variance to the variation in genotypes. Numerous models have been proposed to incorporate additive genetic effects into models for prediction or association. However, there is a scarcity of models that can adequately account for gene by gene or other forms of genetical interactions. In addition, there is an increased interest in using marker annotations in genome-wide prediction and association. In this paper, we discuss an hybrid modeling methodology which combines the parametric mixed modeling approach and the non-parametric rule ensembles. This approach gives us a flexible class of models that can be used to capture additive, locally epistatic genetic effects, gene x background interactions and allows us to incorporate one or more annotations into the genomic selection or association models. We use benchmark data sets covering a range of organisms and traits in addition to simulated data sets to illustrate the strengths of this approach. The improvement of model accuracies and association results suggest that a part of the "missing heritability" in complex traits can be captured by modeling local epistasis.

Existing approaches to combine both additive and multiplicative neural units either use a fixed assignment of operations or require discrete optimization to determine what function a neuron should perform. This leads either to an inefficient distribution of computational resources or an extensive increase in the computational complexity of the training procedure. We present a novel, parameterizable transfer function based on the mathematical concept of non-integer functional iteration that allows the operation each neuron performs to be smoothly and, most importantly, differentiablely adjusted between addition and multiplication. This allows the decision between addition and multiplication to be integrated into the standard backpropagation training procedure.

And the artificial intelligence chatbot that didn't survive a day on the Internet. Plus: Reviewing Santiago Calatrava's latest, how to be unprofessional and the "Grand Theft Auto" modification that may have you watching for hours on end. Time has Russian drone footage that provides an overview of what remains of the old Silk Road crossroads, as well as the contemporary human settlement of Tadmur that sits nearby. About 80% of the artifacts appear to be largely intact. The country's antiquities chief says repairs will take five years.

Eric Thayer/GettyHonda Motors demonstrates its Asimo robot during a media preview of the 2014 New York International Auto Show. Bank of America Merrill Lynch recently came out with its "Transforming World Atlas" research note, which examines global economic trends through a series of maps. One notable map showed which countries had the highest number of operational robots. Japan was number one with 310,508 operational robots, according to data from 2012. There's even a hotel staffed almost entirely by robots that opened last year in Nagasaki, Japan, according to BAML.