Governments and international organizations the world over are investing towards the goal of achieving universal energy access for improving socio-economic development. However, in developing settings, monitoring electrification efforts is typically inaccurate, infrequent, and expensive. In this work, we develop and present techniques for high-resolution monitoring of electrification progress at scale. Specifically, our 3 unique contributions are: (i) identifying areas with(out) electricity access, (ii) quantifying the extent of electrification in electrified areas (percentage/number of electrified structures), and (iii) differentiating between customer types in electrified regions (estimating the percentage/number of residential/non-residential electrified structures). We combine high-resolution 50 cm daytime satellite images with Convolutional Neural Networks (CNNs) to train a series of classification and regression models. We evaluate our models using unique ground truth datasets on building locations, building types (residential/non-residential), and building electrification status. Our classification models show a 92% accuracy in identifying electrified regions, 85% accuracy in estimating percent of (low/high) electrified buildings within the region, and 69% accuracy in differentiating between (low/high) percentage of electrified residential buildings. Our regressions show $R^2$ scores of 78% and 80% in estimating the number of electrified buildings and number of residential electrified building in images respectively. We also demonstrate the generalizability of our models in never-before-seen regions to assess their potential for consistent and high-resolution measurements of electrification in emerging economies, and conclude by highlighting opportunities for improvement.

This demo paper introduces the BLAB Reporter, a robot-journalist covering the Brazilian Blue Amazon. The Reporter is based on a pipeline architecture for Natural Language Generation; it offers daily reports, news summaries and curious facts in Brazilian Portuguese. By collecting, storing and analysing structured data from publicly available sources, the robot-journalist uses domain knowledge to generate and publish texts in Twitter. Code and corpus are publicly available

Large, monolithic datacentres at the heart of enterprises could give way to hundreds or thousands of smaller data stores and devices, each with their own storage capacity. This driver for this is organisations moving their processes to the business "edge". Edge computing is no longer simply about putting some local storage into a remote or branch office (ROBO). Rather, it is being driven by the internet of things (IoT), smart devices and sensors, and technologies such as autonomous cars. All these technologies increasingly need their own local edge data storage.

Today's world is full of sensors, and the higher your nation-state is on the advanced-industrial food chain, the more likely it is that you carry a sensor on your person every minute of every day (and for many, even while asleep). That matters: the data collected by these sensors can be stored, analyzed, and weaponized. And although most of today's data collectors are for-profit corporations, there are dire risks alongside the potential for breakthroughs in areas such as medicine and global warming. The collection, analysis, storage, and theft of information about you have lethal implications; both for you as an individual and for all of us in terms of interstate war. In his 2018 book AI Superpowers, author and entrepreneur Kai-Fu Lee likened big data to the new crude oil and noted that insofar as the analogy holds, that would make the People's Republic of China (PRC) the world's data Saudi Arabia.

Decades ago, a civil war in Sierra Leone left thousands as amputees. Researcher and current Education Minister David Moinina Sengeh set out to help them with a more comfortable socket for prostheses. David Moinina Sengeh is a biomechatronics engineer and the current Minister of Education and Chief Innovation Officer in his home country of Sierra Leone. He pioneered a new system for creating prosthetic sockets, which fit a prosthetic leg onto a patient's residual limb. Using multiple technologies, Sengeh created sockets that are far more comfortable than traditional ones, and can be produced cheaply and quickly.

Growing demand for healthcare services, rising shortage of clinical resources, unequal access and unwarranted variation in care have contributed to high interest in the role of artificial intelligence (AI) in the healthcare sector. AI applications are being developed and piloted across the entire spectrum of the industry ranging from clinical diagnostics, treatment procedures (surgical robotics), personal health applications, population stratification and pharmaceutical research to hospital administration and workflows. However, most of the AI applications under development are quite narrow in their focus, having been designed for very specific tasks/decisions, using standardised data from limited sample sets. These AI applications inevitably fail when exposed to the real world. The challenges of unstructured, non-standardised data, diverse patient populations, varying processes and treatment protocols need to be resolved to enhance the potential of AI in healthcare. When it comes to successes of AI in healthcare, diagnostic imaging has so far shown the most promise. All of the 120 plus AI algorithms that have been approved by US Food and Drug Administration (as of June 2021) for use are related to diagnostic imaging. AI applications are increasingly becoming mainstream across radiology departments. Drug discovery has been another promising area with numerous companies witnessing varying levels of success. Moderna famously used AI technologies to speed up vaccine discovery and development for Covid-19. AI technologies developed by Google have been slowly replacing conventional drug discovery methods. Population health management, hospital operations management, personal health diagnostics are some other areas where AI is making inroads in the healthcare sector. Increasing penetration of wearables combined with the growing adoption of mobile health apps has the potential of putting the power of managing health in the hands of the consumer. The Middle East as a market presents a unique set of challenges and opportunities that need to be highlighted and addressed in the coming years if it means to fully unlock this potential. It will require an integrated approach across four primary dimensions, driven by the key stakeholders in the healthcare sector.

A large number of ancient mosaics have not reached us because they have been destroyed by erosion, earthquakes, looting or even used as materials in newer construction. To make things worse, among the small fraction of mosaics that we have been able to recover, many are damaged or incomplete. Therefore, restoration and reconstruction of mosaics play a fundamental role to preserve cultural heritage and to understand the role of mosaics in ancient cultures. This reconstruction has traditionally been done manually and more recently using computer graphics programs but always by humans. In the last years, Artificial Intelligence (AI) has made impressive progress in the generation of images from text descriptions and reference images. State of the art AI tools such as DALL-E2 can generate high quality images from text prompts and can take a reference image to guide the process. In august 2022, DALL-E2 launched a new feature called outpainting that takes as input an incomplete image and a text prompt and then generates a complete image filling the missing parts. In this paper, we explore whether this innovative technology can be used to reconstruct mosaics with missing parts. Hence a set of ancient mosaics have been used and reconstructed using DALL-E2; results are promising showing that AI is able to interpret the key features of the mosaics and is able to produce reconstructions that capture the essence of the scene. However, in some cases AI fails to reproduce some details, geometric forms or introduces elements that are not consistent with the rest of the mosaic. This suggests that as AI image generation technology matures in the next few years, it could be a valuable tool for mosaic reconstruction going forward.

We consider quantitative analyses of spectral data using laser-induced breakdown spectroscopy. We address the small size of training data available, and the validation of the predictions during inference on unknown data. For the purpose, we build robust calibration models using deep convolutional multitask learning architectures to predict the concentration of the analyte, alongside additional spectral information as auxiliary outputs. These secondary predictions can be used to validate the trustworthiness of the model by taking advantage of the mutual dependencies of the parameters of the multitask neural networks. Due to the experimental lack of training samples, we introduce a simulation-based data augmentation process to synthesise an arbitrary number of spectra, statistically representative of the experimental data. Given the nature of the deep learning model, no dimensionality reduction or data selection processes are required. The procedure is an end-to-end pipeline including the process of synthetic data augmentation, the construction of a suitable robust, homoscedastic, deep learning model, and the validation of its predictions. In the article, we compare the performance of the multitask model with traditional univariate and multivariate analyses, to highlight the separate contributions of each element introduced in the process.

When subjected to automated decision-making, decision subjects may strategically modify their observable features in ways they believe will maximize their chances of receiving a favorable decision. In many practical situations, the underlying assessment rule is deliberately kept secret to avoid gaming and maintain competitive advantage. The resulting opacity forces the decision subjects to rely on incomplete information when making strategic feature modifications. We capture such settings as a game of Bayesian persuasion, in which the decision maker offers a form of recourse to the decision subject by providing them with an action recommendation (or signal) to incentivize them to modify their features in desirable ways. We show that when using persuasion, the decision maker and decision subject are never worse off in expectation, while the decision maker can be significantly better off. While the decision maker's problem of finding the optimal Bayesian incentive-compatible (BIC) signaling policy takes the form of optimization over infinitely-many variables, we show that this optimization can be cast as a linear program over finitely-many regions of the space of possible assessment rules. While this reformulation simplifies the problem dramatically, solving the linear program requires reasoning about exponentially-many variables, even in relatively simple cases. Motivated by this observation, we provide a polynomial-time approximation scheme that recovers a near-optimal signaling policy. Finally, our numerical simulations on semi-synthetic data empirically demonstrate the benefits of using persuasion in the algorithmic recourse setting.

The proliferation of DeepFake technology is a rising challenge in today's society, owing to more powerful and accessible generation methods. To counter this, the research community has developed detectors of ever-increasing accuracy. However, the ability to explain the decisions of such models to users is lacking behind and is considered an accessory in large-scale benchmarks, despite being a crucial requirement for the correct deployment of automated tools for content moderation. We attribute the issue to the reliance on qualitative comparisons and the lack of established metrics. We describe a simple set of metrics to evaluate the visual quality and informativeness of explanations of video DeepFake classifiers from a human-centric perspective. With these metrics, we compare common approaches to improve explanation quality and discuss their effect on both classification and explanation performance on the recent DFDC and DFD datasets.