To what extent have RL researchers tapped into Experimental Design theory from statisticians? They have focused on minimizing costs of data collection that optimizes predictability / explainability for all sorts of problems.

We view the creative arts as an industry indisputably connected to human ability. Writing a novel, connecting the dots in a piece of investigative journalism, painting a landscape, filming and editing a movie -- these are all creative pursuits that require the human touch. But is it possible that robots could learn to do these things too? Could creative jobs also be on the chopping block? One category where AI is currently being pushed to its creative limits is content marketing.

We consider the task of learning the parameters of a {\em single} component of a mixture model, for the case when we are given {\em side information} about that component, we call this the "search problem" in mixture models. We would like to solve this with computational and sample complexity lower than solving the overall original problem, where one learns parameters of all components. Our main contributions are the development of a simple but general model for the notion of side information, and a corresponding simple matrix-based algorithm for solving the search problem in this general setting. We then specialize this model and algorithm to four common scenarios: Gaussian mixture models, LDA topic models, subspace clustering, and mixed linear regression. For each one of these we show that if (and only if) the side information is informative, we obtain parameter estimates with greater accuracy, and also improved computation complexity than existing moment based mixture model algorithms (e.g. tensor methods). We also illustrate several natural ways one can obtain such side information, for specific problem instances. Our experiments on real data sets (NY Times, Yelp, BSDS500) further demonstrate the practicality of our algorithms showing significant improvement in runtime and accuracy.

The way you control Apple's smart speaker, the HomePod, is pretty cool. You just talk to it. But Siri, Apple's voice assistant, can't make the speaker do much beyond the basics and can only summon music from Apple-owned cloud services. That sets it behind the other products in the voice-activated speaker race--most notably those from Amazon or Google--which are capable, flexible, and largely platform-neutral when it comes to streaming. It's too bad Siri can't play internet radio or the deep cuts from your local MP3 library, because the speaker itself sounds absolutely amazing.

Several Facebook groups are dedicated to it, one with over 3,000 likes, New Yorker readers overwhelmingly selected it as the word to eliminate from the dictionary, and Jimmy Fallon sarcastically thanked it for being the worst word in the English language. When you ask people why this might be, there is no shortage of armchair theory: that there's something about the sounds involved, that it puts your face in a position similar to the facial expression of disgust, or that it reminds people of mold or sex. Psychologist Paul Thibodeau and his colleagues ran a study to sort it out. They gave participants a set of words and asked them to rate how, whether, and to what degree, each word made them uncomfortable. Twenty-one percent of the people in the study had an aversion to the unloved word.

Artificial intelligence is moving into movie production. Screenwriters denied the big budgets and formidable resources of the major film studios may soon have another option, thanks to a new algorithm that can generate a video simply by consuming a (very short) script. The new movies are far from Oscar-worthy, but a similar technique could one day find uses outside entertainment, by, say, helping a witness reconstruct a car crash or a crime. Artificial intelligence (AI) is getting much better at identifying the content of images and providing labels. So-called "generative" algorithms go the other way, producing images from labels (or brain scans). A few can even take a single movie frame and predict the next series of frames.

I watched the movie Black Panther this past week, and it was the first time in a movie when I had a hard time deciphering what was real and not real. The ships looked ultra-realistic, the costumes so convincing and utterly detailed that you almost think you could buy one at Nordstrom on sale this weekend. One of the most digitally detailed and colorful movies ever made, it's also a warning about the future. Not knowing when something is actually just a robot. This was not the case just a year or two ago.

Once large law firms had armies of first-year law graduates, combing documents for relevant information; now machines largely do it. New artificial intelligence diagnosed lung cancer 50 percent more accurately than radiology experts last year. And the U.S. Postal Service plans to deploy autonomous trucks by 2025. These are signs of big change, precipitated by a wave of new artificial intelligence resulting from a perfect storm of investments and development these past five years. And coming developments will increasingly enable machines to do more mental and physical tasks faster, better and cheaper than humans.

If you're a music fan, the first wave of smart speakers was probably a disappointment. While Alexa and Google Assistant have definitively proven they have a place in the home, the first Echo and Google Home devices were unimpressive when it came to actually playing music. They did the job in a pinch, and being able to command Spotify with your voice is a killer feature, but many longed for better-quality audio. Fortunately, that call has been answered. In the past six months, Sonos, Google and Apple have all released music-first speakers with voice assistants built in. If you value audio quality above all else (and have about $400 to spend), what's the right smart speaker for you?

Star Wars droid characters BB8, R2D2 and C3P0 make an appearance at the Academy Awards. Ever since Star Wars first hit theaters in the '70s, many fans have wondered what it might be like to have your own personal R2-D2 by your side. A researcher at Texas A&M says while modern robots aren't exactly the droids we're looking for right now, there are signs robots use similar techniques as R2-D2 and BB-8, introduced during the latest Star Wars movie trilogy. "The word'droid' has become so ubiquitous in our collective consciousness that it is hard to believe that the word was created, and trade marked, by George Lucas," wrote Robin R. Murphy in the latest issue of Science Robotics. "But, it is believable that Star Wars robots can imitate or motivate real science." A feature of both droids applicable to today's robots is the use of nonverbal communication, Murphy writes.