Across the globe, energy systems are changing and creating unprecedented challenges for the organisations tasked with ensuring the lights stay on. In the UK, National Grid is facing shrinking margins, looming capacity shortages and unpredictable peaks and troughs in energy supply caused by increasing levels of renewable penetration. At the Reinventing Energy Summit, Michael Bironneau, Head of Technology Development at Open Energi, will explore how the same machine learning techniques that have let machines defeat chess and Go masters, can also be leveraged to orchestrate massive amounts of flexible demand-side capacity – from industrial equipment, co-generation and battery storage systems – towards the one goal of creating a smarter grid; one that is cleaner, cheaper, more secure and more efficient. For World Cities Day 2016, I asked Michael a few questions to learn more about utilising data science in energy, creating a smarter grid, political challenges, and more. What are the main transformative technologies that will help create a smarter grid?

Amazon's Echo voice-controlled smart speaker is finally available in the UK, but was the wait worth it? The Echo is one of the first devices with Amazon's voice assistant - a rival to Apple's Siri, Google's Assistant and Microsoft's Cortana - which allows you to control music playback just by speaking to it and a whole lot more. Echo is three devices in one. It's a voice-controlled Wi-Fi and Bluetooth speaker capable of playing music from Amazon music, a Spotify premium account or a smartphone, tablet or computer connected via Bluetooth. It's also a smart voice assistant called Alexa that's capable of answering queries, setting timers, doing calculations, telling you the weather or what's in your calendar and other bits you might expect from Siri, Cortana or Google Assistant.

Dutch carrier KLM has begun using artificial intelligence (AI) to better serve the many customers who communicate with the airline via social media. "Applying AI, KLM can handle a greater volume of questions while still maintaining its personal approach and speed," Tjalling Smit, senior vice president, digital at Air France-KLM, said recently when announcing the technology. The carrier's move comes as flyers around the world increasingly use services such as Twitter, Facebook Messenger, WeChat and Instagram to submit comments, complaints and questions to airlines. Carriers have had to adapt, employing teams of customer service agents to respond to postings both public and private. KLM has been a leader in that area.

TOKYO SoftBank Group Chairman and CEO Masayoshi Son showed me a photo on his iPhone and said, "I will never forget this scene for the rest of my life." The photo showed a group of white yachts in a bay under an endless blue sky in Marmaris, a port town in southern Turkey. Son swiped the screen, and a selfie photo of him in chino pants and a casual shirt appeared. Several hours before the photos were taken on July 4, Son met Simon Segars, CEO of ARM Holdings, and Stuart Chambers, chairman of the British computer chip design company, on the second floor of a restaurant overlooking the bay. Chambers had arrived in Marmaris, a popular resort town, after receiving an unexpected phone call from Son while yachting with his family in the Mediterranean Sea.

Existing methods for structure discovery in time series data construct interpretable, compositional kernels for Gaussian process regression models. While the learned Gaussian process model provides posterior mean and variance estimates, typically the structure is learned via a greedy optimization procedure. This restricts the space of possible solutions and leads to over-confident uncertainty estimates. We introduce a fully Bayesian approach, inferring a full posterior over structures, which more reliably captures the uncertainty of the model.

Over the years, ensemble methods have become a staple of machine learning. Similarly, generalized linear models (GLMs) have become very popular for a wide variety of statistical inference tasks. The former have been shown to enhance out- of-sample predictive power and the latter possess easy interpretability. Recently, ensembles of GLMs have been proposed as a possibility. On the downside, this approach loses the interpretability that GLMs possess. We show that minimum description length (MDL)-motivated compression of the inferred ensembles can be used to recover interpretability without much, if any, downside to performance and illustrate on a number of standard classification data sets.

Adaptive stochastic gradient methods such as AdaGrad have gained popularity in particular for training deep neural networks. The most commonly used and studied variant maintains a diagonal matrix approximation to second order information by accumulating past gradients which are used to tune the step size adaptively. In certain situations the full-matrix variant of AdaGrad is expected to attain better performance, however in high dimensions it is computationally impractical. We present Ada-LR and RadaGrad two computationally efficient approximations to full-matrix AdaGrad based on randomized dimensionality reduction. They are able to capture dependencies between features and achieve similar performance to full-matrix AdaGrad but at a much smaller computational cost. We show that the regret of Ada-LR is close to the regret of full-matrix AdaGrad which can have an up-to exponentially smaller dependence on the dimension than the diagonal variant. Empirically, we show that Ada-LR and RadaGrad perform similarly to full-matrix AdaGrad. On the task of training convolutional neural networks as well as recurrent neural networks, RadaGrad achieves faster convergence than diagonal AdaGrad.

Random Forests (RFs) are strong machine learning tools for classification and regression. However, they remain supervised algorithms, and no extension of RFs to the one-class setting has been proposed, except for techniques based on second-class sampling. This work fills this gap by proposing a natural methodology to extend standard splitting criteria to the one-class setting, structurally generalizing RFs to one-class classification. An extensive benchmark of seven state-of-the-art anomaly detection algorithms is also presented. This empirically demonstrates the relevance of our approach.

Convolutional networks have marked their place over the last few years as the best performing model for various visual tasks. They are, however, most suited for supervised learning from large amounts of labeled data. Previous attempts have been made to use unlabeled data to improve model performance by applying unsupervised techniques. These attempts require different architectures and training methods. In this work we present a novel approach for unsupervised training of Convolutional networks that is based on contrasting between spatial regions within images. This criterion can be employed within conventional neural networks and trained using standard techniques such as SGD and back-propagation, thus complementing supervised methods.

Through machine learning, researchers have repeated the historic criminology experiment of telling criminals apart from law-abiding people using facial recognition. Physiognomy, the ability to judge a person's character from appearance alone, has been around since ancient Greece and was widely accepted by philosophers. Although the theory has generally been disbanded, studies still crop up now and again. Xiaolin Wu and Xi Zhang, Chinese researchers from Shanghai Jiao Tong University, released a controversial paper on arXiv, an online open-sourced pre-print journal – it has not been published officially. They have singled out three features that can supposedly tell if a person is more likely to be a delinquent or not by probing upper lip curvature, eye inner corner distance, and the angle from nose tip to two mouth corners (nose-mouth angle). It's bad news for those who have smaller mouths, curvier upper lips and closer-set eyes, as you look more like a crook, apparently.