Thank you for your help building out the list!]

Those who dream of getting an Amazon package, a prescription drug, or even a beer delivered to their doorsteps via drone might have their wishes fulfilled sooner than expected.

Wiraya Activation Intelligence will use machine learning to help mobile operators serve millions of customers per day across Europe.

In July, China unveiled a plan to become the world leader in artificial intelligence and create an industry worth $150 billion to its economy by 2030.

UK-bsased ASI Data Science recently unveiled a new machine learning algorithm capable of identifying jihadist content with a staggering 94 percent accuracy. In London, reporters were given a first-hand look at the inner workings of the algorithm, though they were asked not to share the actual methodology. According to BBC's Dave Lee, the "algorithm draws on characteristics typical of [The Islamic State] and its online activity." From what we can piece together, the algorithm appears to use image recognition to examine videos and determine the similarity to other, confirmed videos of the same nature. After thousands of hours of video training, it begins to spot patterns and unique characteristics it can apply to videos outside its training dataset.

NASA's sun-observing spacecraft has captured colorful images of the moment the earth blocked its view of the sun. In a short animation posted on Tuesday, the Solar Dynamics Observatory's (SDO) view of a purple-colored sun is interrupted as Earth completely covers its surface. The sun isn't actually purple, but it looks that color because the images were taken in a wavelength of extreme ultraviolet light. This kind of ultraviolet light is a type that's usually invisible to the human eye, so NASA noted that it's been colorized in purple. Stunning footage from NASA's Solar Dynamics Observatory captures when the earth completely covers its view of the sun.

New insight has emerged on the contents of Elon Musk's Falcon Heavy, which blasted off from Cape Canaveral earlier this month. The Tesla Roadster is carrying a small disc developed by researchers at the University of Southampton in England. The disc, which looks like a shrunken DVD, is storing information on human knowledge and was designed by a group aiming to preserve history called the Arch Mission Foundation. This includes Isaac Asimov's Foundation trilogy, a series of science fiction books. A disc made by the Arch Mission Foundation is on board Elon Musk's orbiting Tesla Roadster.

Submodular functions have applications throughout machine learning, but in many settings, we do not have direct access to the underlying function $f$. We focus on stochastic functions that are given as an expectation of functions over a distribution $P$. In practice, we often have only a limited set of samples $f_i$ from $P$. The standard approach indirectly optimizes $f$ by maximizing the sum of $f_i$. However, this ignores generalization to the true (unknown) distribution. In this paper, we achieve better performance on the actual underlying function $f$ by directly optimizing a combination of bias and variance. Algorithmically, we accomplish this by showing how to carry out distributionally robust optimization (DRO) for submodular functions, providing efficient algorithms backed by theoretical guarantees which leverage several novel contributions to the general theory of DRO. We also show compelling empirical evidence that DRO improves generalization to the unknown stochastic submodular function.

We present a framework to learn privacy-preserving encodings of images (or other high-dimensional data) to inhibit inference of a chosen private attribute. Rather than encoding a fixed dataset or inhibiting a fixed estimator, we aim to to learn an encoding function such that even after this function is fixed, an estimator with knowledge of the encoding is unable to learn to accurately predict the private attribute, when generalizing beyond a training set. We formulate this as adversarial optimization of an encoding function against a classifier for the private attribute, with both modeled as deep neural networks. We describe an optimization approach which successfully yields an encoder that permanently limits inference of the private attribute, while preserving either a generic notion of information, or the estimation of a different, desired, attribute. We experimentally validate the efficacy of our approach on private tasks of real-world complexity, by learning to prevent detection of scene classes from the Places-365 dataset.

We demonstrate that state-of-the-art optical character recognition (OCR) based on deep learning is vulnerable to adversarial images. Minor modifications to images of printed text, which do not change the meaning of the text to a human reader, cause the OCR system to "recognize" a different text where certain words chosen by the adversary are replaced by their semantic opposites. This completely changes the meaning of the output produced by the OCR system and by the NLP applications that use OCR for preprocessing their inputs.