The capsules of Capsule Networks are collections of neurons that represent an object or part of an object in a parse tree. The output vector of a capsule encodes the so called instantiation parameters of this object (e.g. position, size, or orientation). The routing-by-agreement algorithm routes output vectors from lower level capsules to upper level capsules. This iterative algorithm selects the most appropriate parent capsule so that the active capsules in the network represent nodes in a parse tree. This parse tree represents the hierarchical composition of objects out of smaller and smaller components. In this paper, we will show experimentally that the routing-by-agreement algorithm does not ensure the emergence of a parse tree in the network. To ensure that all active capsules form a parse tree, we introduce a new routing algorithm called dynamic deep routing. We show that this routing algorithm allows the training of deeper capsule networks and is also more robust to white box adversarial attacks than the original routing algorithm.

Even for a company as serially scandalous as Facebook, it's been a bad week for the social network. Separate investigations revealed that Facebook gave more than 150 firms access to people's private messages, while also making it impossible for users to avoid location-based ads. After months of fallout from the Cambridge Analytica scandal, US prosecutors also finally got around to filing a lawsuit against Facebook for its data sharing practices. Individually, none of these would likely be enough to bring Facebook down, but some experts believe that, collectively, this could signal the end for the internet behemoth. David Carroll, an associate professor at Parsons School of Design in New York, said this week may finally have dealt Facebook its "knockout" blow.

In 2018, artificial intelligence (AI) researchers made breakthroughs in accelerating machine learning training, anticipating cybersecurity attacks, and eliminating bias. The year 2019 promises to take society from today's "narrow" AI to a new era of "broad" AI, where developers, enterprises, and consumers will be able to fully take advantage of the technology's potential, according to a Thursday blog post from Dario Gil, COO and vice president of AI and quantum at IBM Research. "Broad AI will be characterized by the ability to learn and reason more broadly across tasks, to integrate information from multiple modalities and domains, all while being more explainable, secure, fair, auditable and scalable," Gil wrote in the post. SEE: IT leader's guide to the future of artificial intelligence (Tech Pro Research) Most of today's AI methods are fundamentally based on correlations, and lack a deep understanding of causality, Gil wrote. "Emerging causal inference methods allow us to infer causal structures from data, to efficiently select interventions to test putative causal relationships, and to make better decisions by leveraging knowledge of causal structure," he added.

LONDON - British police say two people were arrested early Saturday morning for suspected "criminal use of drones'" in the Gatwick Airport case that has created nightmarish travel delays for tens of thousands of holiday passengers. Sussex police did not release the age or gender of the two suspects arrested late Friday night and did not say where the arrests were made. The two have not been charged. Police Superintendent James Collis asked the public in the Gatwick area to remain vigilant. "Our investigations are still ongoing, and our activities at the airport continue to build resilience to detect and mitigate further incursions from drones by deploying a range of tactics," he said.

Heart failure (HF) is one of the leading causes of hospital admissions in the US. Readmission within 30 days after a HF hospitalization is both a recognized indicator for disease progression and a source of considerable financial burden to the healthcare system. Consequently, the identification of patients at risk for readmission is a key step in improving disease management and patient outcome. In this work, we used a large administrative claims dataset to (1)explore the systematic application of neural network-based models versus logistic regression for predicting 30 days all-cause readmission after discharge from a HF admission, and (2)to examine the additive value of patients' hospitalization timelines on prediction performance. Based on data from 272,778 (49% female) patients with a mean (SD) age of 73 years (14) and 343,328 HF admissions (67% of total admissions), we trained and tested our predictive readmission models following a stratified 5-fold cross-validation scheme. Among the deep learning approaches, a recurrent neural network (RNN) combined with conditional random fields (CRF) model (RNNCRF) achieved the best performance in readmission prediction with 0.642 AUC (95% CI, 0.640-0.645). Other models, such as those based on RNN, convolutional neural networks and CRF alone had lower performance, with a non-timeline based model (MLP) performing worst. A competitive model based on logistic regression with LASSO achieved a performance of 0.643 AUC (95%CI, 0.640-0.646). We conclude that data from patient timelines improve 30 day readmission prediction for neural network-based models, that a logistic regression with LASSO has equal performance to the best neural network model and that the use of administrative data result in competitive performance compared to published approaches based on richer clinical datasets.

Although science and the public celebrated the amazing predictive power of the new machine learningmethods, many researchers are left with some unease, simply because prediction does not imply causation. The ability to uncover causal relations is, however, at the core of most questions concerning the effects of particular policies, medical treatments, marketing campaigns, businessdecisions, etc. (see Athey, 2017, for a recent discussion). The recently rapidly expanding causal machine learning literature holds great promise for the improved estimation of causal effects by merging the statistics and econometrics literature oncausality with the supervised statistical and machine learning (ML) literature focussing on prediction. The classical causality literature clarifies the conditions needed for being able to estimate causal effects. It also shows how to transform a counterfactual causal problem into specific prediction problems (e.g., Imbens and Wooldridge, 2009). The latter literature on ML provides tools that can be highly effective in solving prediction problems (e.g.

NASA has released several stunning new images of Mars captured by the InSight lander's robotic arm as it snapped a photos of its new workspace. A winter wonderland sits amid a sandy Martian surface -- at least, that's the story new images released by the European Space Agency (ESA) from the Red Planet seem to tell. The stunning photos, which reveal a 50-mile-wide crater filled with ice, were shared by the ESA's Mars Express spacecraft on Thursday. The Korolev crater is located on the northern lowlands of Mars, and it's consistently covered in a blanket of ice about a mile thick, the ESA said in a recent news release. "A beautiful #winter wonderland... on #Mars!" the ESA announced in a tweet, which was shared nearly 10,000 times as of Friday afternoon.

Uber's self-driving cars are getting back on the road today, nine months after one of its autonomous vehicles struck and killed a pedestrian in Arizona. The firm said it's resuming limited testing on public roads in Pittsburgh. The return of testing comes days after the state of Pennsylvania granted Uber permission to resume testing. A fatal accident involving a pedestrian and one of Uber's self-driving vehicles could have been prevented, a new report claims. An employee warned the ride-sharing giant that there were issues with Uber's autonomous-driving technology just days before Elaine Herzberg, a 49-year-old Arizona woman, was struck and killed.

NASA's InSight lander has deployed its first instrument onto the surface of Mars. New images from the lander show the seismometer on the ground, after it was lifted onto the surface by the lander's robotic arm. It will record the waves traveling through the interior structure of the planet, and could help explain mysterious'marsquakes' scientists believe occur regularly. New images from the lander show the seismometer on the ground, after it was lifted onto the surface by the lander's robotic arm. It will record the waves traveling through the interior structure of the planet, and could help explain mysterious'marsquakes' scientists believe occur regularly.

The Army used a cutting-edge Israeli anti-drone system to defeat the unmanned aerial vehicle (UAV) that brought misery to hundreds of thousands of people at Gatwick airport. The British Army bought six'Drone Dome' systems for £15.8 million in 2018 and the technology is used in Syria to destroy ISIS UAVs. Police had been seen on Thursday with an off-the-shelf DJI system that tracks drones made by that manufacturer and shows officers where the operator is (DJI is the most popular commercial drone brand.) However, the drone used at Gatwick is thought to have been either hacked or an advanced non-DJI drone, which rendered the commercial technology used by the police useless. At that point, the Army's'Drone Dome' system made by Rafael was called in.