By combining a deep understanding of plasma physics with machine learning techniques, DIFFER researchers developed a new ultrafast neural network model of the turbulent plasma in a fusion reactor. The neural network can accurately predict heat and particle transport in the fusion reactor up to 100.000 times faster than before: a vital tool to optimize the performance of future fusion power plants. Fusion reactors are fuelled by a plasma: a hot, ionized gas of hydrogen isotopes that fuse together at extreme temperatures to form helium and release clean energy. The behavior of the plasma is not easy to predict: the charged plasma particles respond not only to the magnetic field that keeps them trapped inside the reactor, but also to the electromagnetic fields they create themselves through their own motion. That makes predicting a fusion plasma in order to optimize its state a difficult but rewarding problem to tackle.

Although, artificial intelligence being in trend and a niche has become ubiquitous and in one way or another has touched all domains that a human mind can think about. But there are sectors where AI can play a major role and that discovery is yet to be made. Especially talking about sectors where nature and technology meet, software development industries are still figuring how to transform them keeping in mind the impact on environment. AI in the current scenario is mainly focused on Energy Forecasting, Energy Efficiency and Energy Accessibility, in the renewable energy domain. AI can be a game changer in a lot of ways speaking theoretically.

A U.S. Department of Energy initiative could refurbish existing supercomputers, turning them into high-performance artificial intelligence machines. WASHINGTON, D.C.--The U.S. Department of Energy (DOE) is planning a major initiative to use artificial intelligence (AI) to speed up scientific discoveries. At a meeting here last week, DOE officials said they will likely ask Congress for between $3 billion and $4 billion over 10 years, roughly the amount the agency is spending to build next-generation "exascale" supercomputers. "That's a good starting point," says Earl Joseph, CEO of Hyperion Research, a high-performance computing analysis firm in St. Paul that tracks AI research funding. He notes, though, that DOE's planned spending is modest compared with the feverish investment in AI by China and industry.

In 1985 researchers at the University of Leiden published a paper describing the phase transitions of the heavy fermion alloy uranium ruthenium silicide (URu2Si2). That work sparked numerous studies into this fascinating material, with the phase transition at 17.5K proving particularly puzzling. Despite decades of research, the nature of this phase transition is still unclear. This March a collaboration of researchers from Cornell University, Florida State University, Los Alamos National Laboratory, Max Planck Institute, Dresden, and Leiden University shed further light on the problem by combining resonant ultrasound spectroscopy and machine learning. Their work was published in Science Advances.

Industrial Intelligence was one of the core topics at this year's Hannover Fair. Due to the importance of this area, we took a closer look at the topic area. Artificial intelligence has the potential to revolutionize the production and energy industries. People teach machines to act logically and purposefully to meet customer needs. AI systems generate knowledge and today, based on data and algorithms, can continuously optimize operating states or reliably predict faults and failures – in production processes, in the power grid or in logistics.

The power grid frequency is the central observable in power system control, as it measures the balance of electrical supply and demand. A reliable frequency forecast can facilitate rapid control actions and may thus greatly improve power system stability. Here, we develop a weighted-nearest-neighbor (WNN) predictor to investigate how predictable the frequency trajectories are. Our forecasts for up to one hour are more precise than averaged daily profiles and could increase the efficiency of frequency control actions. Furthermore, we gain an increased understanding of the specific properties of different synchronous areas by interpreting the optimal prediction parameters (number of nearest neighbors, the prediction horizon, etc.) in terms of the physical system. Finally, prediction errors indicate the occurrence of exceptional external perturbations. Overall, we provide a diagnostics tool and an accurate predictor of the power grid frequency time series, allowing better understanding of the underlying dynamics.

Measurements in Liquid Argon Time Projection Chamber (LArTPC) neutrino detectors, such as the MicroBooNE detector at Fermilab, feature large, high fidelity event images. Deep learning techniques have been extremely successful in classification tasks of photographs, but their application to LArTPC event images is challenging, due to the large size of the events. Events in these detectors are typically two orders of magnitude larger than images found in classical challenges, like recognition of handwritten digits contained in the MNIST database or object recognition in the ImageNet database. Ideally, training would occur on many instances of the entire event data, instead of many instances of cropped regions of interest from the event data. However, such efforts lead to extremely long training cycles, which slow down the exploration of new network architectures and hyperparameter scans to improve the classification performance. We present studies of scaling a LArTPC classification problem on multiple architectures, spanning multiple nodes. The studies are carried out on simulated events in the MicroBooNE detector. We emphasize that it is beyond the scope of this study to optimize networks or extract the physics from any results here. Institutional computing at Pacific Northwest National Laboratory and the SummitDev machine at Oak Ridge National Laboratory's Leadership Computing Facility have been used. To our knowledge, this is the first use of state-of-the-art Convolutional Neural Networks for particle physics and their attendant compute techniques onto the DOE Leadership Class Facilities. We expect benefits to accrue particularly to the Deep Underground Neutrino Experiment (DUNE) LArTPC program, the flagship US High Energy Physics (HEP) program for the coming decades.

Artificial Intelligence is moving at the speed of light, with multiple companies creating software, products and services in not just a vertical way – more of a horizontal disruption. Form Healthcare to Security, from Real Estate to Telecom, here is a look into 28 companies powering the disruption with AI – 1st Edition. Sherpa.ai was founded in 2012 after deep research into Artificial Intelligence, with the conviction of creating a personal assistant that would be not just useful, but indispensable for users. In order to do this, Sherpa brought together a team of experts in Artificial Intelligence who, coupled with a fantastic design, have been able to create the next generation of Digital Assistants which will help users make their life not just more exciting, but also more enjoyable. WellSaid Labs has developed state of the art text-to-speech technology that creates life-like synthetic voice, from the voices of real people.

In this letter, we present a novel Gaussian Process Learning-based Probabilistic Optimal Power Flow (GP-POPF) for solving POPF under renewable and load uncertainties of arbitrary distribution. The proposed method relies on a non-parametric Bayesian inference-based uncertainty propagation approach, called Gaussian Process (GP). We also suggest a new type of sensitivity called Subspace-wise Sensitivity, using observations on the interpretability of GP-POPF hyperparameters. The simulation results on 14-bus and 30-bus systems show that the proposed method provides reasonably accurate solutions when compared with Monte-Carlo Simulations (MCS) solutions at different levels of uncertain renewable penetration as well as load uncertainties, while requiring much less number of samples and elapsed time.

Most marketers assume artificial intelligence knows more than it thinks, but it still thinks more than it knows -- and marketers know more than they think. This misconception persists because artificial intelligence (AI) does so much of the tedious work for us. Obtaining AI proficiency as a marketer was much different for me than learning other marketing skills. It's all too easy to believe AI eliminates the need to think critically about what we're doing. In contrast, learning other tools convinced me that mastering new platforms commanded the fullest extent of our brainpower.