On The Tonight Show, in March 1978, the late astronomer Carl Sagan had lots to talk about. He had just published Dragons of Eden: Speculations on the Evolution of Human Intelligence--which would win the Pulitzer Prize--and Star Wars, released the year before, still captivated the public's imagination. When Johnny Carson, the show's then-host, asked Sagan to expand on some comments he'd made prior to the evening, about the film's indifference to scientific accuracy, Sagan said the "11-year-old in me loved" it, but it "could have made a better effort to do things right." His critique would resonate today: After making the biological point that the Star Wars scenario--humans evolving long ago, in a faraway galaxy--is vastly improbable, Sagan said there's another problem: "They're all white." Carson, pushing back a bit, said, "They did have a scene in Star Wars with a lot of strange characters."

The mother of a slain Arizona woman says her daughter had recently revived an online dating profile in search of fairy tale romance, but was instead lured to an apartment by a man now accused of killing her and leaving the body in a shallow desert grave.

Quantum computing's full potential may still be years away, but there are plenty of benefits to be realized right now. So argues Vern Brownell, president and CEO of D-Wave Systems, whose namesake quantum system is already in its second generation. Launched 17 years ago by a team with roots at Canada's University of British Columbia, D-Wave introduced what it called "the world's first commercially available quantum computer" back in 2010. Since then the company has doubled the number of qubits, or quantum bits, in its machines roughly every year. Today, its D-Wave 2X system boasts more than 1,000.

Amazon on Friday boasted that its virtual assistant Alexa is capable of 1,000 "skills," as the online retail giant bolsters defenses against rivals such as Google and Apple. The Seattle-based company said in a blog post that the programs were developed specifically for the voice-commanded Alexa software used in Amazon devices such as Echo and Fire TV. Amazon released voice-enabled wireless speaker Echo in late 2014, infusing it with virtual assistant smarts that enable it to answer questions or control linked devices upon command. A kit lets outside software developers create "experiences," similar in concept to apps, for Alexa. Alexa director Rob Pulciani said tens of thousands of developers are learning about and crafting programs that introduce users "to the magic and simplicity of hands-free, voice-driven interactions."

Switzerland will become the first country in the world to hold a nationwide referendum on the introduction of a basic income on Sunday. The proposal, if passed, would give every adult legally resident in Switzerland an unconditional income of 2,500 Swiss francs ( 1,755; 2,554) a month, whether they work or not. Supporters point to the fact that 21st-Century work is increasingly automated, with more and more traditional jobs, in factories, retail and even in finance and accounting, being done by machines. And they do not need salaries. The campaign has staged some eyecatching demonstrations, including one in which hundreds of "robots" danced through the streets of Zurich, promising to "free" humans from the daily grind of Monday to Friday work, just to pay the bills. "The robots are saying'we don't want to grab your work and make you suffer'," said campaigner Che Wagner.

Machines are becoming more intelligent every year thanks to advances being made by companies like Google, Facebook, Microsoft, and many others. AI agents, as they're sometimes known, can already beat us at complex board games like Go, and they're becoming more competent in a range of other areas. Now a London artificial-intelligence research lab owned by Google has carried out a study to make sure that we can pull the plug on self-learning machines when we want to. DeepMind, bought by Google for a reported 400 million pounds -- about 580 million -- in 2014, teamed up with scientists at the University of Oxford to find a way to make sure that AI agents don't learn to prevent, or seek to prevent, humans from taking control. The paper -- "Safely Interruptible Agents PDF," published on the website of the Machine Intelligence Research Institute (MIRI) -- was written by Laurent Orseau, a research scientist at Google DeepMind, Stuart Armstrong at Oxford University's Future of Humanity Institute, and several others.

Deep Convolutional Neural Networks (CNN) enforce supervised information only at the output layer, and hidden layers are trained by back propagating the prediction error from the output layer without explicit supervision. We propose a supervised feature learning approach, Label Consistent Neural Network, which enforces direct supervision in late hidden layers in a novel way. We associate each neuron in a hidden layer with a particular class label and encourage it to be activated for input signals from the same class. More specifically, we introduce a label consistency regularization called "discriminative representation error" loss for late hidden layers and combine it with classification error loss to build our overall objective function. This label consistency constraint alleviates the common problem of gradient vanishing and tends to faster convergence; it also makes the features derived from late hidden layers discriminative enough for classification even using a simple k-NN classifier, since input signals from the same class will have very similar representations. Experimental results demonstrate that our approach achieves state-of-the-art performances on several public benchmarks for action and object category recognition.

Recent advances in neural variational inference have spawned a renaissance in deep latent variable models. In this paper we introduce a generic variational inference framework for generative and conditional models of text. While traditional variational methods derive an analytic approximation for the intractable distributions over latent variables, here we construct an inference network conditioned on the discrete text input to provide the variational distribution. We validate this framework on two very different text modelling applications, generative document modelling and supervised question answering. Our neural variational document model combines a continuous stochastic document representation with a bag-of-words generative model and achieves the lowest reported perplexities on two standard test corpora. The neural answer selection model employs a stochastic representation layer within an attention mechanism to extract the semantics between a question and answer pair. On two question answering benchmarks this model exceeds all previous published benchmarks.

Estimating the long-term effects of treatments is of interest in many fields. A common challenge in estimating such treatment effects is that long-term outcomes are unobserved in the time frame needed to make policy decisions. One approach to overcome this missing data problem is to analyze treatments effects on an intermediate outcome, often called a statistical surrogate, if it satisfies the condition that treatment and outcome are independent conditional on the statistical surrogate. The validity of the surrogacy condition is often controversial. Here we exploit that fact that in modern datasets, researchers often observe a large number, possibly hundreds or thousands, of intermediate outcomes, thought to lie on or close to the causal chain between the treatment and the long-term outcome of interest. Even if none of the individual proxies satisfies the statistical surrogacy criterion by itself, using multiple proxies can be useful in causal inference. We focus primarily on a setting with two samples, an experimental sample containing data about the treatment indicator and the surrogates and an observational sample containing information about the surrogates and the primary outcome. We state assumptions under which the average treatment effect be identified and estimated with a high-dimensional vector of proxies that collectively satisfy the surrogacy assumption, and derive the bias from violations of the surrogacy assumption, and show that even if the primary outcome is also observed in the experimental sample, there is still information to be gained from using surrogates.

We introduce a family of adaptive estimators on graphs, based on penalizing the $\ell_1$ norm of discrete graph differences. This generalizes the idea of trend filtering [Kim et al. (2009), Tibshirani (2014)], used for univariate nonparametric regression, to graphs. Analogous to the univariate case, graph trend filtering exhibits a level of local adaptivity unmatched by the usual $\ell_2$-based graph smoothers. It is also defined by a convex minimization problem that is readily solved (e.g., by fast ADMM or Newton algorithms). We demonstrate the merits of graph trend filtering through examples and theory.