Infants growing up bilingual have the capability to differentiate between the two languages, even when they are as young as 20 months old, a new study published Tuesday by Princeton University found. Researchers at Princeton Baby Lab studied how babies and young children learn to see, talk and comprehend the world. International researchers teamed up with researchers from Princeton University and found infants as young as 20 months of age could accurately and efficiently process two languages separately. "By 20 months, bilingual babies already know something about the differences between words in their two languages," Casey Lew-Williams, an assistant professor of psychology and co-director of the Princeton Baby Lab, said. "They do not think that'dog' and'chien' [French] are just two versions of the same thing," he added.

A Foxconn employee has allegedly taken photos of the much-anticipated iPhone 8 that reveal the phone's inner workings. The images were posted to the social media site Weibo and appear to depict an iPhone 8 in the'engineering validation test' (EVT) stage of manufacturing. The images appear to show a wireless charging coil, giving fuel to the rumors of wireless charging. The image depicts a wireless charging coil, front-facing camera, sensors and a stacked logic board design. The empty space on the bottom and right-hand side is most likely for a large L-shaped battery.

Imagine that tomorrow, some smart kid invented a technology that let people or physical goods pass through walls, and posted instructions for how to build it cheaply from common household materials. Lots of industries would probably become more productive. Being able to walk through walls instead of being forced to use doors would make it easier to navigate offices, move goods in and out of warehouses and accomplish any number of mundane tasks. That would give the economy a boost. But the negative might well outweigh the positive.

Machine Learning is one of the most exciting fields in the world. Every week we discover something new, something amazing, something revolutionary. That's why we created This Week in Machine Learning! Each week we publish a curated list of Machine Learning stories as a resource to help you keep pace with all these exciting developments. New posts will be published here first, and previous posts are archived on the Udacity blog.

We approach structured output prediction by optimizing a deep value network (DVN) to precisely estimate the task loss on different output configurations for a given input. Once the model is trained, we perform inference by gradient descent on the continuous relaxations of the output variables to find outputs with promising scores from the value network. When applied to image segmentation, the value network takes an image and a segmentation mask as inputs and predicts a scalar estimating the intersection over union between the input and ground truth masks. For multi-label classification, the DVN's objective is to correctly predict the F1 score for any potential label configuration. The DVN framework achieves the state-of-the-art results on multi-label prediction and image segmentation benchmarks.

We propose a method for learning Markov network structures for continuous data without invoking any assumptions about the distribution of the variables. The method makes use of previous work on a non-parametric estimator for mutual information which is used to create a non-parametric test for multivariate conditional independence. This independence test is then combined with an efficient constraint-based algorithm for learning the graph structure. The performance of the method is evaluated on several synthetic data sets and it is shown to learn considerably more accurate structures than competing methods when the dependencies between the variables involve non-linearities.

Circular variables arise in a multitude of data-modelling contexts ranging from robotics to the social sciences, but they have been largely overlooked by the machine learning community. This paper partially redresses this imbalance by extending some standard probabilistic modelling tools to the circular domain. First we introduce a new multivariate distribution over circular variables, called the multivariate Generalised von Mises (mGvM) distribution. This distribution can be constructed by restricting and renormalising a general multivariate Gaussian distribution to the unit hyper-torus. Previously proposed multivariate circular distributions are shown to be special cases of this construction. Second, we introduce a new probabilistic model for circular regression, that is inspired by Gaussian Processes, and a method for probabilistic principal component analysis with circular hidden variables. These models can leverage standard modelling tools (e.g. covariance functions and methods for automatic relevance determination). Third, we show that the posterior distribution in these models is a mGvM distribution which enables development of an efficient variational free-energy scheme for performing approximate inference and approximate maximum-likelihood learning.

Software is eating the world and Artificial Intelligence (AI) is at the heart of this takeover. Since we at Asgard are deeply involved in the European AI market, we thought we would share our insights. The United Kingdom takes the lead as the strongest AI ecosystem in Europe. We have counted 121 AI firms in the UK, with London clearly the largest hub. In second place is Germany (51), with Berlin as the main AI hub supporting 30 AI companies.

The US Army has increasingly used small consumer drones in the field, purchasing them as needed from consumer manufacturers like the well-known Chinese maker DJI. But documents indicate that the Army Aviation Directorate is now enforcing new orders, banning DJI drones "due to increased awareness of cyber vulnerabilities associated with DJI products." The documents, first obtained by Small UAS News, don't explain the Army's security concern, but refer to classified studies about DJI drones that first went out at the end of May. Previously, hackers have been able to jailbreak some DJI drones to control and modify things like safety features on the devices. Some reports have also indicated that DJI can gather location, audio, and even visual data from user flights. It's unclear what data DJI can access without customer consent, but location and media data from an Army drone could potentially reveal extensive information about US military operations.

At The Seeing Eye guide dog facility in Morristown, New Jersey, puppies spend the first three weeks of their life in plastic kiddie pools--the blue kind, speckled with colorful undersea cartoons. Lined with fleece and towels, the wading pools are perfectly sized for a litter of soft, tiny puppies and their mama to lounge and nurse. And some of those floppy-eared, tiny-tailed puppies will eventually grow into responsible, diligent, working dogs. But to don that Seeing Eye guide dog vest, a dog needs to be exceptional. Guide dogs need to be healthy, unfazed by jackhammers, and smart enough to lead a human across the street only when the coast is clear.