In response, the government takes some scripted measures like shutting down schools on occasion and enforcing the infamous Odd-Even scheme, forcing half the cities cars off the streets. Pollution in Delhi has several causes: seasonal stubble burning in neighboring states, vehicular emission, as well as smoke from power plants and brick kilns dotting the national capital region. Fighting pollution needs a multi-pronged approach -- government policies alone are not enough, they need to be coupled with action on the ground. Let's take brick kilns as an example. Before we can address the pollution caused by them, we need to know exactly how many such kilns are there and their location, whether they are increasing in number or decreasing, and how many are adopting technology to reduce emission, as mandated by the law. Satellite imagery coupled with deep learning can answer these questions, increase accountability and drive results on the ground.

Geo-data firm Fugro collects and analyses information about the Earth and the structures built upon it. It surveys the land and in the case of mapping objects on the sea floor, Fugro uses side scan sonar, collected via boats, to gather information. One project sees Fugro search the sea for boulders to help its customers determine whether they can set up an offshore windfarm. "Windfarm companies want to know where the impediments and where the potential sites they can build windfarms are," Fugro senior innovation engineer Marcus Nepveaux said, speaking at AWS re:Invent in Las Vegas. "So we go in, we map the sea floor for them, tell them where the big rocks or the little rocks are … they may be as small as a foot, and as big as we can detect."

Sophisticated machine learning applications require not only enormous amounts of training data, but powerful computer hardware on which to train. An analysis conducted by San Francisco research firm OpenAI found that since 2012, the amount of compute used in the largest training runs has been increasing exponentially with a 3.4-month doubling time, and that it's grown by more than 300,000 times over that same time period. The trend spurred the development of supercomputers like the U.S. Department of Energy's Sierra and Summit, which leverage dedicated accelerator chips to speed up AI computation. Now, IBM's Hardware Center, in collaboration with New York State, SUNY Polytechnic Institute, and other members of IBM's AI Hardware Center, has delivered a new machine for the Department of Computer Science at Rensselaer Polytechnic Institute (RPI) that's optimized for state-of-the-art machine learning workloads. It's dubbed Artificial Intelligence Multiprocessing Optimized System, or AiMOS (in honor of Rensselaer cofounder Amos Eaton), and it will principally tackle projects in biology, chemistry, the humanities, and related domains underway at the new IBM Research AI Hardware Center on the SUNY campus in Albany.

Businesses in the oil and gas arena have been able to use a wide variety of technological products and solutions. Along with oil and gas software, the field has started using drones, electronic monitoring, internet of things (IoT), data analysis tools, and artificial intelligence. All of these origins of technology have become a discovery in the oil and gas industry in recent years. With the combination of this technology, many oil and gas companies have been able to reach their composition goals and improve their overall facilities management services. Technology, alongside oil and gas software solutions, has been able to help companies increase their safety measures.

Just behind us, a giant industrial magnet powered up with warning signs dotted about its perimeter so we wouldn't scramble our phones. Before long, John Durrell, a specialist in superconductor engineering (who took apart more machines as a teenager than he can remember), arrived with a set of tools in his hands and a glint in his eye.

Rapid advancements in the Internet of Things (IoT) have facilitated more efficient deployment of smart environment solutions for specific user requirement. With the increase in the number of IoT devices, it has become difficult for the user to control or operate every individual smart device into achieving some desired goal like optimized power consumption, scheduled appliance running time, etc. Furthermore, existing solutions to automatically adapt the IoT devices are not capable enough to incorporate the user behavior. This paper presents a novel approach to accurately configure IoT devices while achieving the twin objectives of energy optimization along with conforming to user preferences. Our work comprises of unsupervised clustering of devices' data to find the states of operation for each device, followed by probabilistically analyzing user behavior to determine their preferred states. Eventually, we deploy an online reinforcement learning (RL) agent to find the best device settings automatically. Results for three different smart homes' data-sets show the effectiveness of our methodology. To the best of our knowledge, this is the first time that a practical approach has been adopted to achieve the above mentioned objectives without any human interaction within the system.

In 2018, the number of Australian households with rooftop solar passed 2 million – that's one in five.¹ Tomorrow's smart grid will be a constellation of many generation sources working together, shifting from the traditional one-way power flows from generation through grids to consumers to two-way flows including from the customers back into the grid. As we move towards decentralisation, there is an urgent need for new business models and the technology to support it. A new wave of innovative technologies such as Internet of Things (IoT), Edge and Fog computing, blockchain, machine learning and Artificial Intelligence (AI) will become key enablers for such a transformation. Cisco in partnership with the University of Technology Sydney (UTS) and SAS embarked on a trial where the feasibility and economic benefits of DER aggregation and a real-time energy brokerage in a residential framework were successfully designed, tested and verified.

You are invited to prepare an extended abstract to be considered for presentation at the 2020 Oil & Gas HPC Conference hosted by the Ken Kennedy Institute at Rice University. The conference is the premier meeting place for HPC users and participants to engage in conversations about challenges and opportunities in high performance computing, computational science and engineering, and data science across the energy industry. Attended by more than 500 leaders and experts from the energy industry, academia, national labs, and IT industry, this is a unique annual opportunity for key stakeholders to engage and network to help advance HPC in the industry. Computation, data, and information technology continue to stand out across the energy industry as critical business enablers. Recent advances in machine learning, deep learning, robotics and AI are emerging, and there is convergence between these emerging areas and HPC. With the end of Moore's law, challenges are mounting around a rapidly changing technology landscape. However, the end of one era is also an opportunity for advancements and the beginning of a new era – a renaissance for system architectures highlights the need for investments in people (workforce), algorithms, software innovations, and hardware platforms to support system scalability and demands for increasing digitization across the oil and gas sector.

In BO, nonparametric Gaussian processes (GPs) provide flexible and fast-to-evaluate surrogates of the objective functions. Sequential design decisions, so-called acquisitions, judiciously balance exploration and exploitation in search for global optima, leveraging the uncertainty estimates provided by the GP posterior distributions (see Mockus et al. (1978); Jones et al. (1998) for early works or Shahriari et al. (2015) for a recent review). One of the weaknesses of vanilla BO lies in the underlying assumption that the objective function is a realisation of a GP: when this assumption is strongly violated, the GP model is weakly predictive and BO becomes inefficient. Two classical examples where BO fails are ill-conditioned problems, when the objective function has strong variations on the domain boundaries but is very flat in its central region (or conversely), and non-Lipschitz objectives, for instance with local discontinuities. High conditioning is typical in "exploratory" optimisation problems, when the parameter space is initially chosen very large. Discontinuities are frequent in computational fluid dynamics problems for instance, where a small change in the parameters results in a change of physics (e.g.

Data assimilation for parameter and state estimation in subsurface transport problems remains a significant challenge due to the sparsity of measurements, the heterogeneity of porous media, and the high computational cost of forward numerical models. We present a physics-informed deep neural networks (DNNs) machine learning method for estimating space-dependent hydraulic conductivity, hydraulic head, and concentration fields from sparse measurements. In this approach, we employ individual DNNs to approximate the unknown parameters (e.g., hydraulic conductivity) and states (e.g., hydraulic head and concentration) of a physical system, and jointly train these DNNs by minimizing the loss function that consists of the governing equations residuals in addition to the error with respect to measurement data. We apply this approach to assimilate conductivity, hydraulic head, and concentration measurements for joint inversion of the conductivity, hydraulic head, and concentration fields in a steady-state advection--dispersion problem. We study the accuracy of the physics-informed DNN approach with respect to data size, number of variables (conductivity and head versus conductivity, head, and concentration), DNNs size, and DNN initialization during training. We demonstrate that the physics-informed DNNs are significantly more accurate than standard data-driven DNNs when the training set consists of sparse data. We also show that the accuracy of parameter estimation increases as additional variables are inverted jointly.