Spotted: Vienna has installed around 200 pedestrian crossing lights that can recognise when a person wants to cross the road. The system was commissioned by Municipal Department 33 of the City of Vienna and developed by a team at the Institute of Computer Graphics and Vision at TU Graz University. It is intended to replace the push-button system, and can adapt to give large groups and people with disabilities more time to cross. The system uses cameras mounted on the traffic light that have a large visual field. The research team used global movement models and recorded data to develop learning algorithms, which recognise when a pedestrian wants to cross the street.

Companies have to defend against all attacks, while the attackers only have to get through once. And it's about to get much, much worse. The same artificial intelligence technologies used to power speech recognition, self-driving cars, and "deep fake" videos have the potential to be turned to other uses, like creating viruses that morph faster than antivirus companies can keep up, phishing emails that are indistinguishable from real messages written by humans, and intelligently going after a data center's entire perimeter to find the smallest vulnerability and then use it to burrow in. "We already know that a skilled and determined human attacker is the most difficult to catch," said Ryan Shaw, co-founder at Bionic, a Washington, DC-based cybersecurity startup. "However, much like defenders, our adversaries have a scaling problem -- there is only so much time and skill to go around."

WASHINGTON - The Trump administration has moved ahead with a surveillance drone sale to four U.S. allies in the South China Sea region as acting Defense Secretary Patrick Shanahan said Washington will no longer "tiptoe" around Chinese behavior in Asia. The drones would afford greater intelligence gathering capabilities potentially curbing Chinese activity in the region. Shanahan did not directly name China when making accusations of "actors" destabilizing the region in a speech at the annual Shangri-La Dialogue in Singapore on Saturday but went on to say the United States would not ignore Chinese behavior. The Pentagon announced on Friday it would sell 34 ScanEagle drones, made by Boeing Co., to the governments of Malaysia, Indonesia, the Philippines and Vietnam for a total of $47 million. China claims almost all of the strategic South China Sea and frequently lambastes the United States and its allies over naval operations near Chinese-occupied islands.

Apple is offering iPhone users a way to bypass Facebook's and Google's sign-in services when using new apps. That era actually ended quite some time ago. Remember when iPhones were tied to the desktop for updates? So when Apple confirmed Monday that its next desktop operating system upgrade would split up iTunes into three separate apps, for music, TV shows and movies and podcasts, it seemed like an afterthought. "It's a rounding error," something that should have been done a long time ago, says Gene Munster, an analyst and investor with Loup Ventures, iTunes "had gotten way too big."

Amid a slew of updates to iPhones, Macs and iPads, another tech giant took to the stage at Apple's Worldwide Developer Conference to show off the latest version of Minecraft. Microsoft gave WWDC attendees a first look at the new Minecraft Earth augmented reality game, which takes after Pokémon Go to let users create immersive virtual environments in the real world. Thanks to Apple's ARKit, users can build 3D castles, fight off lifelike creepers that sneak up on them and feed virtual chickens through their iPhone. Microsoft first announced Minecraft Earth earlier this year, but the demo during WWDC on Monday marked the first in-depth look at the interactive game. Developers Lydia Winters and Saxs Persson from Mojang, Microsoft's game development studio, came onstage to show how Minecraft Earth works.

We consider optimization of composite objective functions, i.e., of the form $f(x)=g(h(x))$, where $h$ is a black-box derivative-free expensive-to-evaluate function with vector-valued outputs, and $g$ is a cheap-to-evaluate real-valued function. While these problems can be solved with standard Bayesian optimization, we propose a novel approach that exploits the composite structure of the objective function to substantially improve sampling efficiency. Our approach models $h$ using a multi-output Gaussian process and chooses where to sample using the expected improvement evaluated on the implied non-Gaussian posterior on $f$, which we call expected improvement for composite functions (\ei). Although \ei\ cannot be computed in closed form, we provide a novel stochastic gradient estimator that allows its efficient maximization. We also show that our approach is asymptotically consistent, i.e., that it recovers a globally optimal solution as sampling effort grows to infinity, generalizing previous convergence results for classical expected improvement. Numerical experiments show that our approach dramatically outperforms standard Bayesian optimization benchmarks, reducing simple regret by several orders of magnitude.

This paper concerns the problem of learning control policies for an unknown linear dynamical system to minimize a quadratic cost function. We present a method, based on convex optimization, that accomplishes this task robustly: i.e., we minimize the worst-case cost, accounting for system uncertainty given the observed data. The method balances exploitation and exploration, exciting the system in such a way so as to reduce uncertainty in the model parameters to which the worst-case cost is most sensitive. Numerical simulations and application to a hardware-in-the-loop servo-mechanism demonstrate the approach, with appreciable performance and robustness gains over alternative methods observed in both.

Spiking neural networks (SNNs) with a lattice architecture are introduced in this work, combining several desirable properties of SNNs and self-organized maps (SOMs). Networks are trained with biologically motivated, unsupervised learning rules to obtain a self-organized grid of filters via cooperative and competitive excitatory-inhibitory interactions. Several inhibition strategies are developed and tested, such as (i) incrementally increasing inhibition level over the course of network training, and (ii) switching the inhibition level from low to high (two-level) after an initial training segment. During the labeling phase, the spiking activity generated by data with known labels is used to assign neurons to categories of data, which are then used to evaluate the network's classification ability on a held-out set of test data. Several biologically plausible evaluation rules are proposed and compared, including a population-level confidence rating, and an $n$-gram inspired method. The effectiveness of the proposed self-organized learning mechanism is tested using the MNIST benchmark dataset, as well as using images produced by playing the Atari Breakout game.

Cognitive and metacognitive strategy had demonstrated a significant role in self-regulated learning (SRL), and an appropriate use of strategies is beneficial to effective learning or question-solving tasks during a human-computer interaction process. This paper proposes a novel method combining Knowledge Map (KM) based data mining technique with Thinking Map (TM) to detect learner's cognitive and metacognitive strategy in the question-solving scenario. In particular, a graph-based mining algorithm is designed to facilitate our proposed method, which can automatically map cognitive strategy to metacognitive strategy with raising abstraction level, and make the cognitive and metacognitive process viewable, which acts like a reverse engineering engine to explain how a learner thinks when solving a question. Additionally, we develop an online learning environment system for participants to learn and record their behaviors. To corroborate the effectiveness of our approach and algorithm, we conduct experiments recruiting 173 postgraduate and undergraduate students, and they were asked to complete a question-solving task, such as "What are similarities and differences between array and pointer?" from "The C Programming Language" course and "What are similarities and differences between packet switching and circuit switching?" from "Computer Network Principle" course. The mined strategies patterns results are encouraging and supported well our proposed method.

We introduce the first dataset for human edits of machine-generated visual stories and explore how these collected edits may be used for the visual story post-editing task. The dataset, VIST-Edit, includes 14,905 human edited versions of 2,981 machine-generated visual stories. The stories were generated by two state-of-the-art visual storytelling models, each aligned to 5 human-edited versions. We establish baselines for the task, showing how a relatively small set of human edits can be leveraged to boost the performance of large visual storytelling models. We also discuss the weak correlation between automatic evaluation scores and human ratings, motivating the need for new automatic metrics.