This was a banner week for AI (Artificial Intelligence). Well, C3.ai came public and soared on its debut. Keep in mind that C3.ai provides comprehensive software solutions and services for a myriad of large companies, including 3M, Royal Dutch Shell, Raytheon, Baker Hughes and conEdison. "The use of AI and data analytics will become increasingly important in IT as organizations aim to deliver seamless support and predictive capabilities," said Amit Sawhney, who is the Vice President of Services Operations at Dell Technologies. So then, given all the investment and innovation, what might we see next year with AI?

The digital factory provides undoubtedly a great potential for future production systems in terms of efficiency and effectivity. A key aspect on the way to realize the digital copy of a real factory is the understanding of complex indoor environments on the basis of 3D data. In order to generate an accurate factory model including the major components, i.e. building parts, product assets and process details, the 3D data collected during digitalization can be processed with advanced methods of deep learning. In this work, we propose a fully Bayesian and an approximate Bayesian neural network for point cloud segmentation. This allows us to analyze how different ways of estimating uncertainty in these networks improve segmentation results on raw 3D point clouds. We achieve superior model performance for both, the Bayesian and the approximate Bayesian model compared to the frequentist one. This performance difference becomes even more striking when incorporating the networks' uncertainty in their predictions. For evaluation we use the scientific data set S3DIS as well as a data set, which was collected by the authors at a German automotive production plant. The methods proposed in this work lead to more accurate segmentation results and the incorporation of uncertainty information makes this approach especially applicable to safety critical applications.

The growing use of social networks, smartphones that collect and continuously generate data, the growing use of the Internet, the presence of sensors that measure and monitor everything, causes the volume of the produced data is growing exponentially, providing valuable information for society and for companies. All this is Big Data, defined as a large collection of data volume and variety can not be managed with traditional database management tools, but require the use of new technologies and adequate data management systems for storing and analysis, are able to extract their value quickly.[1] With Big Data are experiencing a new revolution, the large amount of data and information available to us, are considered "black gold" of the new millennium. They are fundamental to the predictive analysis and extrapolation of information (Data Mining) developed by research institutes and companies in support of their decision-making strategies. In business intelligence is changing the way to manage information for decision support, they are developing new tools, and down the costs of data collection systems, storage and processing.

Dr. Thomas Berger has landed a NASA grant to research space weather with machine learning. Berger, the executive director of the University of Colorado Boulder Space Weather Technology, Research and Education Center, is leading a team that has received a two-year, $496,000 grant to design a better forecasting system for solar magnetic eruptions on the sun. These events lead to solar flares and coronal mass ejections that can wreak havoc on radio communications, endanger satellites in low Earth orbit, and even destabilize the electric power grid here on Earth. "Up to very recently, there have only been subjective tools, human forecasters who view images of sunspots and use historical data tables to say, 'The probability of this sunspot flaring in the next 24 hours is X%'," Berger said. A 24-hour range for solar eruption forecasts is about the best a human forecaster can do with current technology.

An automated machine learning framework for geospatial data named PAIRS AutoGeo is introduced on IBM PAIRS Geoscope big data and analytics platform. The framework simplifies the development of industrial machine learning solutions leveraging geospatial data to the extent that the user inputs are minimized to merely a text file containing labeled GPS coordinates. PAIRS AutoGeo automatically gathers required data at the location coordinates, assembles the training data, performs quality check, and trains multiple machine learning models for subsequent deployment. The framework is validated using a realistic industrial use case of tree species classification. Open-source tree species data are used as the input to train a random forest classifier and a modified ResNet model for 10-way tree species classification based on aerial imagery, which leads to an accuracy of $59.8\%$ and $81.4\%$, respectively. This use case exemplifies how PAIRS AutoGeo enables users to leverage machine learning without extensive geospatial expertise.

The U.S. Energy Department's research arm is leveraging machine learning technologies to simplify the design process for energy systems ranging from photovoltaics and wind turbines to aircraft engine compressors. The Advanced Research Projects Agency-Energy, or ARPA-E, last month announced 23 research contracts totaling $15 million to incorporate machine learning into energy product designs. The first-phase contracts are part of an ARPA-E initiative dubbed DIFFERENTIATE, standing for--take a breath--Design Intelligence Fostering Formidable Energy Reduction (and) Enabling Novel Totally Impactful Advanced Technology Enhancements. David Tew, an ARPA-E program director, said the two-year machine learning effort is focused on the engineering design process with the goal of optimizing power generation systems. Along with wind turbines and photovoltaics, DIFFERENTIATE also will focus on power conversion and heat transfer systems, aerodynamics, photonics and range of foundational energy technologies.

One perennial concern about new technology is that it will create a more unequal world, reducing wages and widening the divide between the haves and have-nots. Those fears have acquired a more panicked urgency with the rapid adoption of robotics and artificial intelligence. It's easy for this kind of apprehension to lead to a sort of neo-Luddism in which technology is viewed as the enemy. So I want to talk about another possibility: That new technology could actually help reduce inequality. The world should be ready for both outcomes.

Climate change is global, yet its concrete impacts can strongly vary between different locations in the same region. Seasonal weather forecasts currently operate at the mesoscale (> 1 km). For more targeted mitigation and adaptation, modelling impacts to < 100 m is needed. Yet, the relationship between driving variables and Earth's surface at such local scales remains unresolved by current physical models. Large Earth observation datasets now enable us to create machine learning models capable of translating coarse weather information into high-resolution Earth surface forecasts. Here, we define high-resolution Earth surface forecasting as video prediction of satellite imagery conditional on mesoscale weather forecasts. Video prediction has been tackled with deep learning models. Developing such models requires analysis-ready datasets. We introduce EarthNet2021, a new, curated dataset containing target spatio-temporal Sentinel 2 satellite imagery at 20 m resolution, matched with high-resolution topography and mesoscale (1.28 km) weather variables. With over 32000 samples it is suitable for training deep neural networks. Comparing multiple Earth surface forecasts is not trivial. Hence, we define the EarthNetScore, a novel ranking criterion for models forecasting Earth surface reflectance. For model intercomparison we frame EarthNet2021 as a challenge with four tracks based on different test sets. These allow evaluation of model validity and robustness as well as model applicability to extreme events and the complete annual vegetation cycle. In addition to forecasting directly observable weather impacts through satellite-derived vegetation indices, capable Earth surface models will enable downstream applications such as crop yield prediction, forest health assessments, coastline management, or biodiversity monitoring. Find data, code, and how to participate at www.earthnet.tech .

Recent significant developments, such as GPT-3, have been driven by this conjecture. However, as models scale-up, training them efficiently with backpropagation becomes difficult. Because model, pipeline, and data parallelism distribute parameters and gradients over compute nodes, communication is challenging to orchestrate: this is a bottleneck to further scaling. In this work, we argue that alternative training methods can mitigate these issues, and can inform the design of extreme-scale training hardware. Indeed, using a synaptically asymmetric method with a parallelizable backward pass, such as Direct Feedback Alignement, communication needs are drastically reduced. We present a photonic accelerator for Direct Feedback Alignment, able to compute random projections with trillions of parameters. We demonstrate our system on benchmark tasks, using both fully-connected and graph convolutional networks. Our hardware is the first architecture-agnostic photonic co-processor for training neural networks. This is a significant step towards building scalable hardware, able to go beyond backpropagation, and opening new avenues for deep learning.

Machine learning algorithms have been successfully used to approximate nonlinear maps under weak assumptions on the structure and properties of the maps. We present deep neural networks using dense and convolutional layers to solve an inverse problem, where we seek to estimate parameters in a FitzHugh-Nagumo model, which consists of a nonlinear system of ordinary differential equations (ODEs). We employ the neural networks to approximate reconstruction maps for model parameter estimation from observational data, where the data comes from the solution of the ODE and takes the form of a time series representing dynamically spiking membrane potential of a (biological) neuron. We target this dynamical model because of the computational challenges it poses in an inference setting, namely, having a highly nonlinear and nonconvex data misfit term and permitting only weakly informative priors on parameters. These challenges cause traditional optimization to fail and alternative algorithms to exhibit large computational costs. We quantify the predictability of model parameters obtained from the neural networks with statistical metrics and investigate the effects of network architectures and presence of noise in observational data. Our results demonstrate that deep neural networks are capable of very accurately estimating parameters in dynamical models from observational data.