XPONENTIAL 2018, fascinated hundreds of manufacturers, innovators, industry professionals to the Mile High City (Denver) and the event was assembled by the Association for Unmanned Vehicle System International. This "Everything unmanned" show featured, consumer-grade photography drones, weaponized military aircraft, solar-powered submarines and four-legged robots. Lockheed Martin tested the upcoming technology that would allow cargo aircraft to fly autonomously. Various awards were also presented to the outstanding companies. Aeryon Labs, Inc. was awarded the AUVSI XCELLENCE humanitarian award for the deployment of UAV's to aid in aftermath of Hurricane Irma.

On the Sunday morning after the weather cleared, a pair of NASA researchers loaded onto a small plane at the Mammoth Yosemite Airport, a single-runway operation that stretches out before the pyramid peak of Mount Morrison. After final safety checks, the pilots lifted off, marking the Airborne Snow Observatory's inaugural flight of the season. The ASO is a twin-turboprop Beechcraft King Air 90, equipped with a pair of sensors pointing through a glass cutout on the bottom of the plane. The lidar measures the volume of the mountain snowpack while a spectrometer gauges its reflectivity, together providing a highly accurate estimate of how much water will run off the mountain in the spring and when it will flow through California's warren of dams, reservoirs, and aqueducts. The data allows water authorities to more carefully manage the water charging hydroelectric power plants, feeding towns and cities, and nourishing one of the United States' most productive agricultural regions.

We propose a temporally coherent generative model addressing the super-resolution problem for fluid flows. Our work represents a first approach to synthesize four-dimensional physics fields with neural networks. Based on a conditional generative adversarial network that is designed for the inference of three-dimensional volumetric data, our model generates consistent and detailed results by using a novel temporal discriminator, in addition to the commonly used spatial one. Our experiments show that the generator is able to infer more realistic high-resolution details by using additional physical quantities, such as low-resolution velocities or vorticities. Besides improvements in the training process and in the generated outputs, these inputs offer means for artistic control as well. We additionally employ a physics-aware data augmentation step, which is crucial to avoid overfitting and to reduce memory requirements. In this way, our network learns to generate advected quantities with highly detailed, realistic, and temporally coherent features. Our method works instantaneously, using only a single time-step of low-resolution fluid data. We demonstrate the abilities of our method using a variety of complex inputs and applications in two and three dimensions.

We consider the transfer of experience samples (i.e., tuples < s, a, s', r >) in reinforcement learning (RL), collected from a set of source tasks to improve the learning process in a given target task. Most of the related approaches focus on selecting the most relevant source samples for solving the target task, but then all the transferred samples are used without considering anymore the discrepancies between the task models. In this paper, we propose a model-based technique that automatically estimates the relevance (importance weight) of each source sample for solving the target task. In the proposed approach, all the samples are transferred and used by a batch RL algorithm to solve the target task, but their contribution to the learning process is proportional to their importance weight. By extending the results for importance weighting provided in supervised learning literature, we develop a finite-sample analysis of the proposed batch RL algorithm. Furthermore, we empirically compare the proposed algorithm to state-of-the-art approaches, showing that it achieves better learning performance and is very robust to negative transfer, even when some source tasks are significantly different from the target task.

Teaching agents to perform tasks using Reinforcement Learning is no easy feat. As the goal of reinforcement learning agents is to maximize the accumulated reward, they often find loopholes and misspecifications in the reward signal which lead to unwanted behavior. To overcome this, often, regularization is employed through the technique of reward shaping - the agent is provided an additional weighted reward signal, meant to lead it towards a desired behavior. The weight is considered as a hyper-parameter and is selected through trial and error, a time consuming and computationally intensive task. In this work, we present a novel multi-timescale approach for constrained policy optimization, called, 'Reward Constrained Policy Optimization' (RCPO), which enables policy regularization without the use of reward shaping. We prove the convergence of our approach and provide empirical evidence of its ability to train constraint satisfying policies.

Dot product operations between matrices are at the heart of almost any field in science and technology. In many cases, they are the component that requires the highest computational resources during execution. For instance, deep neural networks such as VGG-16 require up to 15 giga-operations in order to perform the dot products present in a single forward pass, which results in significant energy consumption and thus limits their use in resource-limited environments, e.g., on embedded devices or smartphones. One common approach to reduce the complexity of the inference is to prune and quantize the weight matrices of the neural network and to efficiently represent them using sparse matrix data structures. However, since there is no guarantee that the weight matrices exhibit significant sparsity after quantization, the sparse format may be suboptimal. In this paper we present new efficient data structures for representing matrices with low entropy statistics and show that these formats are especially suitable for representing neural networks. Alike sparse matrix data structures, these formats exploit the statistical properties of the data in order to reduce the size and execution complexity. Moreover, we show that the proposed data structures can not only be regarded as a generalization of sparse formats, but are also more energy and time efficient under practically relevant assumptions. Finally, we test the storage requirements and execution performance of the proposed formats on compressed neural networks and compare them to dense and sparse representations. We experimentally show that we are able to attain up to x15 compression ratios, x1.7 speed ups and x20 energy savings when we lossless convert state-of-the-art networks such as AlexNet, VGG-16, ResNet152 and DenseNet into the new data structures.

New technologies are generating far more information than ever before to help scientists assess and predict the health and behavior of species and ecosystems, as well as the threats they face. These include cryptic cameras, acoustic sensors, satellite imagery and citizen science apps. Now, researchers and conservation practitioners analyzing large data sets are exploring artificial intelligence, or AI--the ability of a machine or a computer program to think and learn--to help them process, analyze and interpret data to monitor ecosystems and predict results. Computer systems already exist that can host huge amounts of data, use AI with increasingly "smart" algorithms to classify data from the various types of sensors used by scientists, apply modeling results to create reproducible code, and create user interfaces to allow people to monitor natural systems and make predictions with high accuracy. By training computer algorithms with a subset of available data, machines can now learn what they should do for a given challenge--such as classify photos by the species found in them, identify areas of a satellite image containing water or intact forest, or translate speech from one language to another --based on human feedback and data collected from previous experience.

Artificial intelligence and energy production might not be so unrelated, according to a faculty member at the University of Wyoming. At the Wyoming Energy Summit on Wednesday, Nick Cheney -- a visiting assistant professor in UW's computer science department -- discussed some of the ways in which autonomous machines using artificial intelligence might change the energy sector in years to come. "I'm not a futurist," Cheney said. "I don't believe, despite working in AI, that artificial intelligence is a thing that will immediately save us all or doom us all. It comes with its own potential risks and benefits.

Taking a look back at another week of news from Cupertino, this week's Apple Loop includes the new iPhone designs for 2018, production starting on the A12 chips, Apple issuing credit notes for battery replacements, a big decision over the MacBook Air, thoughts on WWDC's announcements, Razer's macOS graphics enclosure, Apple working with VW, and an important date for macOS. Apple Loop is here to remind you of a few of the very many discussions that have happened around Apple over the last seven days (and you can read my weekly digest of Android news here on Forbes). Apple is expected to announce three new iPhone handsets in September, and the one that is catching the most attention right now is the'replacement' for the iPhone SE - all the features of the iPhone X but in a much smaller form factor. Gordon Kelly reports on the latest leaks around the design of the super small smartphone: So what do we learn? That Olixar is convinced Apple will fulfill the dreams of millions of iPhone fans, by releasing a new iPhone X variant which crams a bezel-less front display and Face ID technology into a chassis the same size as the iPhone SE. Backing this up, Mobile Fun has passed me new CAD designs showing the new iPhone's dimensions as 121.04 x 55.82 mm (4.8 x 2.2-inches) - fractionally smaller than the 123.8 x 58.6 mm (4.87 x 2.31-inches) of the iPhone SE.

This post was done in partnership with Wirecutter. When readers choose to buy Wirecutter's independently chosen editorial picks, it may earn affiliate commissions that support its work. After six summers of researching, testing, and recommending window air conditioners, we've learned that quiet and affordable ACs make most people the happiest--and we think the LG LW8016ER will fit the bill in most rooms. This 8,000 Btu unit cools as efficiently and effectively as any model with an equal Btu rating, and runs at a lower volume and deeper pitch than others at this price. Little extra features like a fresh-air vent, two-axis fan blades, and a removable drain plug help set it apart, too. The LG LW8016ERis a top choice for an office or den, and some people will find it quiet enough for a bedroom, too. If our main pick is sold out, grab the Frigidaire FFRE0833S1. The Frigidaire is also a little bit easier to install because it's smaller and lighter. If you're buying an air conditioner for your bedroom and don't mind paying a little extra, treat yourself to the Frigidaire Gallery FGRQ08L3T1. It's the quietest window AC we've tested over the past few years. It's also easy to install, and it comes with plenty of extra foam for insulation. By June 2017, this model had sold out for the season, so you may need to act fast if you're interested in it this year. We're also recommending the Frigidaire FGRQ0833U1 as a backup in case the L3T1 goes out of stock. The 33U1 is more expensive, and we haven't tested it ourselves, but Frigidaire confirmed to us that the two models are essentially the same. Both are 8,000 Btu rated, have Energy Efficiency Ratios of 12, and are designed to cool rooms up to 350 square feet.