Integro-difference equation (IDE) models describe the conditional dependence between the spatial process at a future time point and the process at the present time point through an integral operator. Nonlinearity or temporal dependence in the dynamics is often captured by allowing the operator parameters to vary temporally, or by re-fitting a model with a temporally-invariant linear operator at each time point in a sliding window. Both procedures tend to be excellent for prediction purposes over small time horizons, but are generally time-consuming and, crucially, do not provide a global prior model for the temporally-varying dynamics that is realistic. Here, we tackle these two issues by using a deep convolution neural network (CNN) in a hierarchical statistical IDE framework, where the CNN is designed to extract process dynamics from the process' most recent behaviour. Once the CNN is fitted, probabilistic forecasting can be done extremely quickly online using an ensemble Kalman filter with no requirement for repeated parameter estimation. We conduct an experiment where we train the model using 13 years of daily sea-surface temperature data in the North Atlantic Ocean. Forecasts are seen to be accurate and calibrated. A key advantage of our approach is that the CNN provides a global prior model for the dynamics that is realistic, interpretable, and computationally efficient. We show the versatility of the approach by successfully producing 10-minute nowcasts of weather radar reflectivities in Sydney using the same model that was trained on daily sea-surface temperature data in the North Atlantic Ocean.

New generation geostationary satellites make solar reflectance observations available at a continental scale with unprecedented spatiotemporal resolution and spectral range. Generating quality land monitoring products requires correction of the effects of atmospheric scattering and absorption, which vary in time and space according to geometry and atmospheric composition. Many atmospheric radiative transfer models, including that of Multi-Angle Implementation of Atmospheric Correction (MAIAC), are too computationally complex to be run in real time, and rely on precomputed look-up tables. Additionally, uncertainty in measurements and models for remote sensing receives insufficient attention, in part due to the difficulty of obtaining sufficient ground measurements. In this paper, we present an adaptation of Bayesian Deep Learning (BDL) to emulation of the MAIAC atmospheric correction algorithm. Emulation approaches learn a statistical model as an efficient approximation of a physical model, while machine learning methods have demonstrated performance in extracting spatial features and learning complex, nonlinear mappings. We demonstrate stable surface reflectance retrieval by emulation (R2 between MAIAC and emulator SR are 0.63, 0.75, 0.86, 0.84, 0.95, and 0.91 for Blue, Green, Red, NIR, SWIR1, and SWIR2 bands, respectively), accurate cloud detection (86\%), and well-calibrated, geolocated uncertainty estimates. Our results support BDL-based emulation as an accurate and efficient (up to 6x speedup) method for approximation atmospheric correction, where built-in uncertainty estimates stand to open new opportunities for model assessment and support informed use of SR-derived quantities in multiple domains.

Autonomous agents must often deal with conflicting requirements, such as completing tasks using the least amount of time/energy, learning multiple tasks, or dealing with multiple opponents. In the context of reinforcement learning~(RL), these problems are addressed by (i)~designing a reward function that simultaneously describes all requirements or (ii)~combining modular value functions that encode them individually. Though effective, these methods have critical downsides. Designing good reward functions that balance different objectives is challenging, especially as the number of objectives grows. Moreover, implicit interference between goals may lead to performance plateaus as they compete for resources, particularly when training on-policy. Similarly, selecting parameters to combine value functions is at least as hard as designing an all-encompassing reward, given that the effect of their values on the overall policy is not straightforward. The later is generally addressed by formulating the conflicting requirements as a constrained RL problem and solved using Primal-Dual methods. These algorithms are in general not guaranteed to converge to the optimal solution since the problem is not convex. This work provides theoretical support to these approaches by establishing that despite its non-convexity, this problem has zero duality gap, i.e., it can be solved exactly in the dual domain, where it becomes convex. Finally, we show this result basically holds if the policy is described by a good parametrization~(e.g., neural networks) and we connect this result with primal-dual algorithms present in the literature and we establish the convergence to the optimal solution.

Convolutional neural networks (CNNs) have been increasingly deployed to edge devices. Hence, many efforts have been made towards efficient CNN inference on resource-constrained platforms. This paper attempts to explore an orthogonal direction: how to conduct more energy-efficient training of CNNs, so as to enable on-device training? We strive to reduce the energy cost during training, by dropping unnecessary computations, from three complementary levels: stochastic mini-batch dropping on the data level; selective layer update on the model level; and sign prediction for low-cost, low-precision back-propagation, on the algorithm level. Extensive simulations and ablation studies, with real energy measurements from an FPGA board, confirm the superiority of our proposed strategies and demonstrate remarkable energy savings for training. For example, when training ResNet-74 on CIFAR-10, we achieve aggressive energy savings of >90% and >60%, while incurring a top-1 accuracy loss of only about 2% and 1.2%, respectively. When training ResNet-110 on CIFAR-100, an over 84% training energy saving is achieved without degrading inference accuracy.

David Obst a,b, Badih Ghattas b, Sandra Claudel a, Jairo Cugliari c, Yannig Goude a, Georges Oppenheim d a EDF R&D, Palaiseau, France b Institut de Math ematiques de Marseille, Aix-Marseille Universit e, France c ERIC, Universit e de Lyon 2, France d Laboratoire d'Analyse et de Math ematiques Appliqu ees Universit e Paris-Est, Champs-sur-Marne, FranceAbstract While ubiquitous, textual sources of information such as company reports, social media posts, etc. are hardly included in prediction algorithms for time series, despite the relevant information they may contain. In this work, openly accessible daily weather reports from France and the United-Kingdom are leveraged to predict time series of national electricity consumption, average temperature and wind-speed with a single pipeline. Two methods of numerical representation of text are considered, namely traditional Term Frequency - Inverse Document Frequency (TF-IDF) as well as our own neural word embedding. Using exclusively text, we are able to predict the aforementioned time series with sufficient accuracy to be used to replace missing data. Furthermore the proposed word embeddings display geometric properties relating to the behavior of the time series and context similarity between words. Introduction Whether it is in the field of energy, finance or meteorology, accurately predicting the behavior of time series is nowadays of paramount importance for optimal decision making or profit. While the field of time series forecasting is extremely prolific from a research point-of-view, up to now it has narrowed its efforts on the exploitation of regular numerical features extracted from sensors, data bases or stock exchanges. Unstructured data such as text on the other hand remains underexploited for prediction tasks, despite its potentially valuable informative content. Empirical studies have already proven that textual sources such as news articles or blog entries can be correlated to stock exchange time series and have explanatory power for their variations [1, 2]. This observation has motivated multiple extensive experiments to extract relevant features from textual documents in different ways and use them for prediction, notably in the field of finance. In Lavrenko et al. [3], language models (considering only the presence of a word) are used to estimate the probability of trends such as surges or falls of 127 different stock values using articles from Biz Yahoo!. Their results show that this text driven approach could be used to make profit on the market. One of the most conventional ways for text representation is the TF-IDF (Term Frequency - Inverse Document Frequency) approach.

Artificial intelligence is becoming the future of everything. Yet, only a few large companies have the talent and the technology to perfect it. That's the gist of New York Times story published late last week. Rising costs for AI research are locking out university researchers and garage entrepreneurs, two of the traditional -- and historically best -- founts of innovation. In the past, software engineers used code to build platforms and new business models.

NEWPORT NEWS, VA – From new particle accelerator technology, to the exploration of new ways to treat wastewater, to applications of artificial intelligence, six cutting-edge projects are getting a jumpstart on research and development at the Department of Energy's Thomas Jefferson National Accelerator Facility. The projects are supported by the Laboratory Directed Research and Development program, which recently announced the continuation of four projects and new funding for two more for fiscal year 2020, which began October 1. The LDRD program provides resources for Jefferson Lab personnel to make rapid and significant contributions to critical science and technology problems that further the goals of the laboratory and the DOE. "We are delighted with the progress that was made on the ongoing LDRD projects, and we look forward with great interest to the results of the fiscal year 2020 projects and the boost they will give to long-term strategic directions of the laboratory," said Jefferson Lab Director Stuart Henderson. Of the six funded projects, four include aspects of artificial intelligence and machine learning: Three projects aim to develop machine learning to assist physicists in monitoring and/or analyzing large volumes of scientific data, while the last has the goal of improving up-time of Jefferson Lab's Continuous Electron Beam Accelerator Facility, a DOE User Facility.

WASHINGTON, D.C. – Today, the U.S. Department of Energy (DOE) announced $13 million in funding for five research projects in artificial intelligence (AI) aimed at improving AI as a tool of scientific investigation and prediction. The projects involve scientists at nine research institutions, including both DOE national laboratories and universities. "Artificial intelligence, including machine learning, provides an extremely powerful way of tackling the most pressing issues facing our scientists today," said U.S. Secretary of Energy Rick Perry. "This research will help us adapt AI to the specific scientific challenges that DOE-supported scientists are addressing today and in the process help sustain U.S. leadership in this critical and growing field." Of the $13 million, $11.1 million is reserved for two three-year projects focused on the development of new AI algorithms and software adapted to specific scientific problems.

Indian Institute of Technology (IIT) Madras researchers have developed algorithms that enable novel applications for Artificial Intelligence (AI), Machine Learning (ML) and Deep Learning to solve engineering problems. The Researchers are going to establish a startup to deploy their AI Software called'AISoft' to develop solutions to engineering problems in varied fields such as in thermal management, semiconductors, automobile, aerospace and electronic cooling applications. AI, Machine Learning and Deep Learning are now being used for over a decade but traditionally only in areas such as signal processing, speech recognition, image reconstruction and prediction. Very limited attempts have been made globally in using these algorithms in solving engineering problems such as thermal management, electronic cooling industries, automobile problems like fluid dynamics prediction over a bonnet or inside the engine, aerospace industries like aerodynamics and fluid dynamics problems across an aero-foil or turbine engine. A team of researchers lead by Dr. Vishal Nandigana, Assistant Professor, Fluid Systems Laboratory, Department of Mechanical Engineering, IIT Madras, has developed AI and Deep Learning algorithms to solve engineering problems, which they do not solve a physical law to arrive at the solution of the system.

Oral-B has launched its Genius X toothbrush which uses artificial intelligence to help you brush your teeth better for US$220. The Oral-B 10000 Genius X is available from their website for US$220 is the follow up to the Genius 9000, which sold from the Shavershop for AU$349. Unfortunately, there is no word on whether the Oral-B Genius X will make its way down under for Christmas. Featuring wireless Bluetooth connection, the Oral-B Genius X links to a dedicated companion app on your phone to time how long you brush your teeth for, how to pressure your applying, where you have been brushing and where you should brush more next time. Utilising sensors within the toothbrush, the device can detect pressure and its location within your mouth, something a reviewer from Forbes was most impressed about. It does this through the "Genius X AI algorithm" which provides a better brush guide, with a full rating as well.