The 2015 ACM A.M. Turing Award recognized work I did 40 years ago, so it is understandable that my interests have changed significantly, with my most recent project being a book, A New Map for Relationships: Creating True Love at Home & Peace on the Planet, co-authored with my wife Dorothie. While, at first glance, the book might seem to have nothing in common with my work on cryptography, my Turing Lecture drew a number of parallels I will bring out in what follows. The story starts in March 1975, when the U.S. National Bureau of Standards (NBS), now known as the National Institute of Standards and Technology (NIST), proposed a Data Encryption Standard (DES) to protect unclassified but sensitive data. Whitfield Diffie, with whom I shared the Award, and I quickly realized that DES's 56-bit key size was inadequate and needed to be increased.

We introduce physics informed neural networks -- neural networks that are trained to solve supervised learning tasks while respecting any given law of physics described by general nonlinear partial differential equations. In this second part of our two-part treatise, we focus on the problem of data-driven discovery of partial differential equations. Depending on whether the available data is scattered in space-time or arranged in fixed temporal snapshots, we introduce two main classes of algorithms, namely continuous time and discrete time models. The effectiveness of our approach is demonstrated using a wide range of benchmark problems in mathematical physics, including conservation laws, incompressible fluid flow, and the propagation of nonlinear shallow-water waves.

Complex architectures of biological neural circuits, such as parallel processing pathways, has been behaviorally implicated in many cognitive studies. However, the theoretical consequences of circuit complexity on neural computation have only been explored in limited cases. Here, we introduce a mechanism by which direct and indirect pathways from cortex to the CA3 region of the hippocampus can balance both contextual gating of memory formation and driving network activity. We implement this concept in a deep artificial neural network by enabling a context-sensitive bias. The motivation for this is to improve performance of a size-constrained network. Using direct knowledge of the superclass information in the CIFAR-100 and Fashion-MNIST datasets, we show a dramatic increase in performance without an increase in network size.

Climate projections continue to be marred by large uncertainties, which originate in processes that need to be parameterized, such as clouds, convection, and ecosystems. But rapid progress is now within reach. New computational tools and methods from data assimilation and machine learning make it possible to integrate global observations and local high-resolution simulations in an Earth system model (ESM) that systematically learns from both. Here we propose a blueprint for such an ESM. We outline how parameterization schemes can learn from global observations and targeted high-resolution simulations, for example, of clouds and convection, through matching low-order statistics between ESMs, observations, and high-resolution simulations. We illustrate learning algorithms for ESMs with a simple dynamical system that shares characteristics of the climate system; and we discuss the opportunities the proposed framework presents and the challenges that remain to realize it.

The difference between the robots of today and all previous forms of automation is that they are so flexible (Editorial, 25 November). Intelligent robots will be utilised in any new enterprise rather than people now because the financial returns are likely to be so much greater, given that there will be no recruitment difficulties, wage demands, overtime claims, strikes, sickness absence, pensions, transport or housing problems to take care of. Factories can be situated anywhere, and HS2 could be redundant before it becomes operational. In the past, workers displaced by automation could rely on new industries springing up to take them on, but in future these will create far more jobs for robots than people across the board. Our whole economic system, which concentrates on profitability and economics rather than the welfare of the population, can only encourage this trend.

Then detection and identification of extreme weather events in large-scale climate simulations is an important problem for risk management, informing governmental policy decisions and advancing our basic understanding of the climate system. Recent work has shown that fully supervised convolutional neural networks (CNNs) can yield acceptable accuracy for classifying well-known types of extreme weather events when large amounts of labeled data are available. However, many different types of spatially localized climate patterns are of interest including hurricanes, extra-tropical cyclones, weather fronts, and blocking events among others. Existing labeled data for these patterns can be incomplete in various ways, such as covering only certain years or geographic areas and having false negatives. This type of climate data therefore poses a number of interesting machine learning challenges. We present a multichannel spatiotemporal CNN architecture for semi-supervised bounding box prediction and exploratory data analysis. We demonstrate that our approach is able to leverage temporal information and unlabeled data to improve the localization of extreme weather events. Further, we explore the representations learned by our model in order to better understand this important data. We present a dataset, ExtremeWeather, to encourage machine learning research in this area and to help facilitate further work in understanding and mitigating the effects of climate change. The dataset is available at extremeweatherdataset.github.io and the code is available at https://github.com/eracah/hur-detect.

With the increasing popularity of machine learning techniques, it has become common to see prediction algorithms operating within some larger process. However, the criteria by which we train these algorithms often differ from the ultimate criteria on which we evaluate them. This paper proposes an end-to-end approach for learning probabilistic machine learning models in a manner that directly captures the ultimate task-based objective for which they will be used, within the context of stochastic programming. We present three experimental evaluations of the proposed approach: a classical inventory stock problem, a real-world electrical grid scheduling task, and a real-world energy storage arbitrage task. We show that the proposed approach can outperform both traditional modeling and purely black-box policy optimization approaches in these applications.

The deep learning framework can effectively identify cracks in nuclear reactors by analyzing individual video frames.

New research has posited that artificial intelligence will increasingly automate operations for the wind and solar industries, boosting their efficiencies in areas such as decision making and planning, condition monitoring, robotics, and inspections. The new position paper published this week by DNV GL -- international accredited registrar and classification society headquartered near Oslo -- entitled Making Renewables Smarter: The benefits, risks, and future of artificial intelligence in solar and wind, outlines the advances being made in robotics, inspections, supply chain, and the way we work and showcases a variety of opportunities for the solar and wind industries to embrace artificial intelligence (AI) applications to improve their efficiency. "The use of artificial intelligence in industries continues at an impressive rate -- in manufacturing, engineering, healthcare, transportation, finance, telecommunications, services, and energy," the authors of the report explain. "Artificial intelligence's ability to use machine learning to analyse historical and new data, make predictions, control physical operations, and make decisions at increasingly higher levels is having an immense impact." The report explores ways in which AI applications like machine learning can impact the efficiency levels of areas involved in the wind and solar industries such as decision making and planning, condition monitoring, robotics, inspections, certifications and supply chain optimization, as well as the way technical work is carried out.

Engineers at Purdue University in Lafayette, Indiana are developing a new system for keeping nuclear reactors safe with artificial intelligence (AI). In the paper published in the IEEE Transactions on Industrial Electronics journal, the researchers introduced a deep learning framework called a naïve Bayes-convolutional neural network that can effectively identify cracks in reactors by analyzing individual video frames. The method could potentially make safety inspections safer. "Regular inspection of nuclear power plant components is important to guarantee safe operations," Mohammad Jahanshahi, an assistant professor at Purdue's Lyles School of Civil Engineering, said in a press release. "However, current practice is time-consuming, tedious, and subjective and involves human technicians reviewing inspection videos to identify cracks in reactors."