We develop and analyze stochastic subgradient methods for optimizing a new, versatile, application-friendly and tractable class of convex risk measures, termed here as mean-semideviations. Their construction relies on on the concept of a risk regularizer, a one-dimensional nonlinear map with certain properties, essentially generalizing the positive part weighting function in the mean-upper-semideviation risk measure. After we formally introduce mean-semideviations, we study their basic properties, and we present a fundamental constructive characterization result, demonstrating their generality. We then introduce and rigorously analyze the MESSAGEp algorithm, an efficient stochastic subgradient procedure for iteratively solving convex mean-semideviation risk-averse problems to optimality. The MESSAGEp algorithm may be derived as an application of the T-SCGD algorithm of (Yang et al., 2018). However, the generic theoretical framework of (Yang et al., 2018) is too narrow and structurally restrictive, as far as optimization of mean-semideviations is concerned, including the classical mean-upper-semideviation risk measure. By exploiting problem structure, we propose a substantially weaker theoretical framework, under which we establish pathwise convergence of the MESSAGEp algorithm, under the same strong sense as in (Yang et al., 2018). The new framework reveals a fundamental trade-off between the smoothness of the random position function and that of the particular mean-semideviation risk measure under consideration. Further, we explicitly show that the class of mean-semideviation problems supported under our framework is strictly larger than the respective class of problems supported in (Yang et al., 2018). Thus, applicability of compositional stochastic optimization is established for a strictly wider spectrum of mean-semideviation problems, justifying the purpose of our work.

Organizations are looking at how AI and IoT can reduce cost, drive efficiencies, and enhance competitive advantage and support emerging business models. It is also clearly observed that some technical innovations from the mainstream of the IT world, or from other industries, are creating opportunities to leverage technology that did not exist previously in the industry. The industry has, in the past, pursued a siloed approach to applications and technologies. This is characterized by the separation of the engineering and operations groups from IT, and the use of stand-alone, best-of-breed applications within the overall scope of IT. As ubiquitous connectivity continues to permeate technology sectors, an increasing need to unite energy technologies, operational technologies (such as sensors and smart devices) and IT (such as big data, advanced analytics and asset performance management [APM]) with consumer technologies (such as social and mobile) is observed in the industry.

As wireless networks evolve towards high mobility and providing better support for connected vehicles, a number of new challenges arise due to the resulting high dynamics in vehicular environments and thus motive rethinking of traditional wireless design methodologies. Future intelligent vehicles, which are at the heart of high mobility networks, are increasingly equipped with multiple advanced onboard sensors and keep generating large volumes of data. Machine learning, as an effective approach to artificial intelligence, can provide a rich set of tools to exploit such data for the benefit of the networks. In this article, we first identify the distinctive characteristics of high mobility vehicular networks and motivate the use of machine learning to address the resulting challenges. After a brief introduction of the major concepts of machine learning, we discuss its applications to learn the dynamics of vehicular networks and make informed decisions to optimize network performance. In particular, we discuss in greater detail the application of reinforcement learning in managing network resources as an alternative to the prevalent optimization approach. Finally, some open issues worth further investigation are highlighted.

Deep learning continues to push state-of-the-art performance for the semantic segmentation of color (i.e., RGB) imagery; however, the lack of annotated data for many remote sensing sensors (i.e. hyperspectral imagery (HSI)) prevents researchers from taking advantage of this recent success. Since generating sensor specific datasets is time intensive and cost prohibitive, remote sensing researchers have embraced deep unsupervised feature extraction. Although these methods have pushed state-of-the-art performance on current HSI benchmarks, many of these tools are not readily accessible to many researchers. In this letter, we introduce a software pipeline, which we call EarthMapper, for the semantic segmentation of non-RGB remote sensing imagery. It includes self-taught spatial-spectral feature extraction, various standard and deep learning classifiers, and undirected graphical models for post-processing. We evaluated EarthMapper on the Indian Pines and Pavia University datasets and have released this code for public use.

Google's mission is to organize the world's information and make it universally accessible and useful. From the start, they have also made significant efforts to do this in a way that doesn't deplete the world's natural resources. The company has been fully carbon neutral since 2007 and ten years later they are hoping they have achieved the next major goal – drawing every watt of energy they use for their business operations from renewable sources. Kate E Brandt, their lead for sustainability, spoke to me about some of the ways they have been tackling this ambitious challenge while she was visiting London to speak at the Economist Sustainability Summit 2018. She told me "We set a goal in 2012 that we wanted to purchase 100% renewable energy for our operations – so it's a longstanding commitment.

Swiss engineering company ABB is considering acquisitions to increase its capabilities in artificial intelligence, CEO Ulrich Spiesshofer said March 29. "We will continually further expand the portfolio of ABB," he said, speaking on the sidelines of the company's annual general meeting in Zurich. This would include investment in organic growth in artificial intelligence (AI), and partnerships with other companies to accelerate areas such as linking AI to industrial robots. ABB, whose products also include charging stations for electric cars and massive converters for continent-spanning transmission systems, would also invest selectively in start-up companies, Spiesshofer said. He was speaking after ABB gave a slightly more upbeat assessment about the development of its markets for 2018, saying conditions had brightened.

Swiss engineering company ABB is considering acquisitions to increase its capabilities in artificial intelligence, CEO Ulrich Spiesshofer said March 29. "We will continually further expand the portfolio of ABB," he said, speaking on the sidelines of the company's annual general meeting in Zurich. This would include investment in organic growth in artificial intelligence (AI), and partnerships with other companies to accelerate areas such as linking AI to industrial robots. ABB, whose products also include charging stations for electric cars and massive converters for continent-spanning transmission systems, would also invest selectively in start-up companies, Spiesshofer said. He was speaking after ABB gave a slightly more upbeat assessment about the development of its markets for 2018, saying conditions had brightened.

The topic of industry disruption -- "a process whereby a smaller company with fewer resources is able to successfully challenge established incumbent businesses" -- is rife with misconceptions. One of the biggest is that it is a mysterious, random, and unpredictable event. Another is that it happens to you in ways that are beyond your control. Those views may have been valid at one time, but they no longer apply. Industry disruption, as Accenture research has found, is reasonably predictable.

Work to scrap the crippled Fukushima No. 1 nuclear plant and deal with radioactive water buildup at the site is expected to cost around ¥220 billion ($2 billion) annually over the three-year period from fiscal 2018, a source said Thursday. It is the first time that Tokyo Electric Power Company Holdings Inc. and the state-backed Nuclear Damage Compensation and Decommissioning Facilitation Corp., or NDF, have provided an estimate of annual costs for cleaning up the Fukushima No. 1 complex, more than seven years after the tsunami-triggered nuclear crisis. Tepco and the NDF will soon submit the financial plan to the government to gain approval from industry minister Hiroshige Seko. The NDF, established after the Fukushima crisis started, holds a majority stake in Tepco, and instructs the utility on how to effectively decommission the plant. The outlay plan comes as total costs to scrap the nuclear plant have ballooned.

Lithium (Li) metal electrodes are not deployable in rechargeable batteries because electrochemical plating and stripping invariably leads to growth of dendrites that reduce coulombic efficiency and eventually short the battery. It is generally accepted that the dendrite problem is exacerbated at high current densities. Here, we report a regime for dendrite evolution in which the reverse is true. In our experiments, we found that when the plating and stripping current density is raised above 9 milliamperes per square centimeter, there is substantial self-heating of the dendrites, which triggers extensive surface migration of Li. We show that repeated doses of high-current-density healing treatment enables the safe cycling of Li-sulfur batteries with high coulombic efficiency.