Which naturally leads to: whose ethical and social values? What biases do we build in – intentionally or otherwise – that will affect the output? Each individual, socio-cultural group, and national geography can have different attitudes to ethics, morality and legality. This is further complicated by the choice of data used to train AI systems. Even the best-laid intentions can go awry if the data itself breaks our ethical standards.

As Labor Day weekend ushers in the unofficial end of summer, Hollywood wraps up a dismal summer box office season with movie ticket sales in the U.S. and Canada the lowest they've been in a decade. Fittingly, the box office ranks remain virtually unchanged from last week with the two films premiering this week, the re-release of Sony's "Close Encounters of the Third Kind" and the Weinstein Co.'s "Tulip Fever" debuting at No. 14 and No. 24, respectively. Lionsgate's R-rated action comedy "The Hitman's Bodyguard" maintained the top spot for the third consecutive week, bringing in an estimated $13.4 million over the holiday weekend for a total of $58 million cumulatively. Starring Samuel L. Jackson as an imprisoned assassin who enlists the help of bodyguard Ryan Reynolds, the $30-million film received mixed reviews from audiences, earning a B rating on CinemaScore but a stale 40% "fresh" rating on Rotten Tomatoes. Warner Bros' horror prequel "Annabelle: Creation" remained at No. 2 on the charts, earning $9.3 million through Monday, for a total of $91 million.

We asked Sebastian Raschka, applied machine learning and deep learning researcher and the author of Packt's best-selling book Python Machine …

Strong interim results posted by Chinese smartphone camera makers may have proven yet again the future of camera manufacturing lies with lenses that are minuscule and A.I.-savvy. The landscape of the camera making industry has fundamentally changed as consumers are shelving their professional cameras and using smartphones to take pictures in recent years. Now as computer vision technology, which uses deep learning scheme to enable machines to understand digital images and videos, matures, smartphone camera makers, many of which are leading the pack in computer vision software development, could displace traditional camera makers' dominance in their own backyard. Companies such as Q Technology, Sunny Optical exemplified the rapid growth of smartphone camera makers in China. Both companies have posted strong first-half results this year, as traditional camera makers such as Canon see a decline in unit sales in digital compact cameras.

Building an experimental city from scratch was always going to be a complex, expensive, problematic task--even Walt would have struggled to pull this one off. But the Walt Disney Co. seems to have abandoned all efforts at serious futurology. Since the abandonment of the "2055" project, the original Tomorrowland--home of Walt's starry-eyed vision of space, atoms, and transit--has displayed little of Walt's futuristic spirit. The park's current "Googie" architecture is a relic from the past, "a space age look that was enormously popular during the 1950s," according to the Disney website. The rides are a mixture of vintage classics (Autopia, the monorail, and the submarine still run) and Pixar/Star Wars–themed attractions that are more science-fiction than science-future.

Guest written by Rebecca Njeri! There are funny mishaps that result from imperfectly trained machine learning algorithms. Or these two guys stuck on a voice activated elevator that doesn't understand their accent. Or maybe Amazon's Alexa trying to order hundreds of dollhousesbecause it confuses the news anchor's report for a request from its owner. There are also the memes on the Amazon Whole Foods purchase, which are truly in the spirit of defective algorithms.

In Poland, a newspaper subscription tends to satisfy purely intellectual needs and is regarded as somewhat of a luxury which the majority of the people can heroically forego; in the United States a newspaper is regarded as a basic need of every person, indispensable as bread itself. The basic need for information, of all kinds, as Mark Twain observed (in Pudd'nhead Wilson, 1897): "The old saw says, 'Let a sleeping dog lie.' Right. Still, when there is much at stake, it is better to get a newspaper to do it." Today, the need for "serious" and "sensational" information is satisfied less and less by newspapers alone. The days of loyalty to a particular news organization on a particular piece of technology in a particular form are gone.

Yes, that is what transfer learning is. Although 8 conv 4 fc is a little unusual. You might be better off finding a pre-trained, tried and tested convnet architecture like VGG or Resnet and save yourself a lot of work. Most DL frameworks provide some kind of "model zoo" to download pre-trained networks.

Back then, it was actually difficult to find datasets for data science and machine learning projects. Since then, we've been flooded with lists and lists of datasets. Today, the problem is not finding datasets, but rather sifting through them to keep the relevant ones. Well, we've done that for you right here. Below, you'll find a curated list of free datasets for data science and machine learning, organized by their use case.