SCI COMMUN### Infectious diseases The 11th Ebola outbreak in the Democratic Republic of the Congo (DRC) is officially over, giving the country respite from the disease for the first time in more than 2 years. On 18 November, the World Health Organization (WHO) announced that no new cases had been identified for 42 days, twice the incubation period for the deadly virus. The outbreak, in the western Équateur province, started in late May, just as a bigger one in the eastern DRC was coming to an end. (That outbreak had killed 2200 people.) The Équateur outbreak sickened 130 and killed 55; a campaign that vaccinated more than 40,000 people is credited with helping end it. Special portable coolers that keep the vaccine at −80°C for up to 1 week allowed health workers to administer the shots in communities deep in the rainforest, accessible only by boat or helicopter. The same technology will be useful in efforts to distribute COVID-19 vaccines in Africa, says Matshidiso Moeti, WHO's regional director. The coronavirus pandemic complicated the fight against Ebola, WHO says, but the expertise gained by local health workers in earlier outbreaks in the region was a major advantage. They will remain on the lookout for potential flare-ups. $1,000,000 —Gift from entertainer Dolly Parton in April to support development of Moderna's coronavirus vaccine, which the company last week said showed an efficacy of 94.5%. “I felt so proud to have been part of that little seed money,” Parton told BBC. ### Marine ecology The Allen Coral Atlas, a project to map the world's shallow coral reefs with high-resolution satellites, last week launched a monitoring system to detect coral bleaching events as they occur. When corals face extreme heat, they expel their algal symbionts, leaving them bone white and vulnerable to stress; repeated bleaching episodes, growing more common with global warming, can cause massive die-offs. The system detects the whitening using imagery from the privately owned Planet satellite constellation, processed with machine learning. A pilot has begun in Hawaii to use the data as an early warning system for researchers, to help them identify and study species both vulnerable and resistant to warming extremes. The monitoring of bleaching is expected to expand next year to shallow reefs globally. ### Diagnostics The U.S. Food and Drug Administration (FDA) issued its first emergency use authorization last week for an at-home diagnostic test that can detect the pandemic coronavirus in just minutes. However, the test might not be widely available until spring 2021. Produced by Lucira Health, a biotech company, it is expected to cost less than $50 and require a doctor's prescription. The company says it will soon distribute tests in parts of California and Florida; it says it needs time to scale up manufacturing for national distribution. Lucira's test amplifies viral genetic material, making it nearly as accurate as laboratory tests that use the polymerase chain reaction, the current gold standard. FDA previously approved at-home tests that must be mailed to a laboratory for analysis. Several other companies are working on rapid antigen tests, which detect viral particles, for home use. But concerns remain about antigen tests' reliability. Still, some public health specialists consider widely available, low-cost, at-home testing vital for controlling the pandemic. ### Funding A new U.S. National Institutes of Health (NIH) award will allow early-career investigators who want to shift research directions when applying for their first independent award to submit a proposal without first generating preliminary data to support their idea. Reviewers will instead assess the soundness of the project's approach. The Katz award is named for Stephen Katz, a longtime champion of young researchers who was director of the National Institute of Arthritis and Musculoskeletal and Skin Diseases when he died in 2018. The grant will build on an NIH policy that prioritizes proposals from early-stage investigators—those no more than 10 years from completing their training who are applying for their first research grant. The policy has been credited with raising their numbers from fewer than 600 supported in 2013 to more than 1300 last year. Applications for the first Katz awards are due on 26 January 2021. ### Leadership Democrats in Congress say a political appointee given a senior post at the U.S. National Institute of Standards and Technology (NIST) is unfit for the job because he lacks technical skills and holds pseudoscientific views about racial differences on IQ tests. On 9 November, Jason Richwine, an independent public policy analyst, took up the new position of deputy undersecretary of commerce for standards and technology, and Commerce Secretary Wilbur Ross subsequently issued an order that would put Richwine in charge of the $1 billion research agency if NIST Director Walter Copan leaves or is fired. On 17 November, Representative Eddie Bernice Johnson (D–TX), who leads the science committee in the U.S. House of Representatives, asked Ross to justify the moves. Richwine has advocated for more restrictive immigration policies, and his 2009 doctoral thesis argued that lower IQ scores by Mexican and Hispanic immigrants suggest a genetic component to intelligence that is “likely to persist over several generations.” ### Diversity The editors of Nature Communications say they are reviewing a paper that drew scalding criticism after it suggested that encouraging female junior scientists to work with female mentors could “hinder the careers of women.” The 17 November study, led by data scientist Bedoor AlShebli of New York University, Abu Dhabi, examined 3 million mentor-protégé pairs and how gender influenced the impact of papers later published by the protégés. Female protégés, it concluded, did better if they worked with male mentors. Critics pounced, noting the authors ignored reviewer complaints about the study's methods and arguing the journal was promoting a harmful and unfounded message. The article's authors said they welcome the review. ### Animal diseases European authorities reported on 19 November they have detected highly pathogenic avian influenza in 302 birds in eight countries. Only 18 cases were in poultry; most of the rest were in wild birds, the European Food Safety Authority and its partners said. The number of infected birds is expected to rise with winter migrations. Several flu strains were identified, but no people were reported to be infected, and the risk of that occurring is considered low; researchers studying the viruses found no genetic markers indicating they had adapted to infect mammals. But the threat to poultry is high, and the report's authors recommended bird producers increase precautions against infections. VACCINE APPLICATION Days after making public the final analysis of their 40,000-person COVID-19 vaccine trial, which found 95% efficacy, Pfizer and its German partner BioNTech filed for emergency authorization of the messenger RNA vaccine from the U.S. Food and Drug Administration—the first such request for a vaccine during the pandemic. They plan to seek additional approvals in other countries soon. Pfizer hopes to supply up to 50 million doses this year. REMDESIVIR PANNED A World Health Organization panel recommended against using the antiviral drug remdesivir to treat most hospitalized COVID-19 patients. Its review of four studies of 7000 people found that the drug, which the U.S. Food and Drug Administration approved last month for hospitalized patients, did not reduce mortality or speed recovery. But the panel encouraged further study of it. AMMO BAN Denmark has become the first nation to ban all lead-based hunting ammunition, including bullets and shotgun pellets, to protect wildlife. Hunters annually release about 2 tons of lead into Denmark's environment; waterbirds and other species eat the toxic material and die. European regulators are considering a ban like Denmark's.

Like it or not, artificial intelligence (AI) is already part of our daily lives. From the smartphones in our pockets to the Alexa virtual assistants on our kitchen counters, AI and its applications are accepted norms today. While we appreciate that AI can automate repetitive workplace tasks or even drive a car, the reality is that its implications are much further reaching. Luminaries like Elon Musk and Bill Gates have spoken out about the potential downsides of AI. At times, they have even issued outright warnings.

On Thursday, 26 November, Prof. Andrew Murray, will deliver the Sixth T.M.C. Asser Lecture – 'Almost Human: Law and Human Agency in the Time of Artificial Intelligence'. Asser Institute researcher Dr. Dimitri Van Den Meerssche had the opportunity to speak with professor Murray about his perspective on the challenges posed by Artificial Intelligence to our human agency and autonomy – the backbone of the modern rule of law. A conversation on algorithmic opacity, the peril of dehumanization, the illusionary ideal of the'human in the loop' and the urgent need to go beyond'ethics' in the international regulation of AI. One central observation in your Lecture is how Artificial Intelligence threatens human agency. Could you elaborate on your understanding of human agency and how it is being threatened? In my Lecture I refer to the definition of agency by legal philosopher Joseph Raz. He argues that to be fully in control of one's own agency and decisions you need to have capacity, the availability of options and the freedom to exercise that choice without interference. My claim is that there are four ways in which the adoption and use of algorithms affect our autonomy, and particularly Raz's third requirement: that we are to be free from coercion. First, there is an internal and positive impact. This happens when an algorithm gives us choices, which have been limited by pre-determined values – values that we cannot observe. The second impact is internal and negative. In this scenario, choices are removed because of pre-selected values.

Mobile money in Africa has rapidly evolved from its traditional role as a payment service to a gateway for millions on the continent to gain access to an ever-increasing array of financial products and services. For banks and other traditional financial service providers, future profitability will greatly depend on their ability to form partnerships with mobile carriers and accurately target subscribers on the network with financial service offerings that are relevant. There is a compelling business argument for effective customer targeting and cross-selling: for banks, digital channels with a high uptake boost low-cost deposit mobilization and increase lending capacity; for mobile carriers, digital financial product offerings that meet subscriber needs deepen engagement and increase retention. In this post, I will explore how machine learning can be used to classify individuals into one of four categories based on the types of financial services they are most likely to use. This is an example of multi-class classification where the task involves using an algorithm to induce a mapping function between a given set of input features and a categorical target variable that takes on more than two values.

What are the opportunities and threats faced by the vendors in the international market?

Future beyond fifth-generation (B5G) and sixth-generation (6G) mobile communications will shift from facilitating interpersonal communications to supporting Internet of Everything (IoE), where intelligent communications with full integration of big data and artificial intelligence (AI) will play an important role in improving network efficiency and providing high-quality service. As a rapid evolving paradigm, the AI-empowered mobile communications demand large amounts of data acquired from real network environment for systematic test and verification. Hence, we build the world's first true-data testbed for 5G/B5G intelligent network (TTIN), which comprises 5G/B5G on-site experimental networks, data acquisition & data warehouse, and AI engine & network optimization. In the TTIN, true network data acquisition, storage, standardization, and analysis are available, which enable system-level online verification of B5G/6G-orientated key technologies and support data-driven network optimization through the closed-loop control mechanism. This paper elaborates on the system architecture and module design of TTIN. Detailed technical specifications and some of the established use cases are also showcased.

While long short-term memory (LSTM) models have demonstrated stellar performance with streamflow predictions, there are major risks in applying these models in contiguous regions with no gauges, or predictions in ungauged regions (PUR) problems. However, softer data such as the flow duration curve (FDC) may be already available from nearby stations, or may become available. Here we demonstrate that sparse FDC data can be migrated and assimilated by an LSTM-based network, via an encoder. A stringent region-based holdout test showed a median Kling-Gupta efficiency (KGE) of 0.62 for a US dataset, substantially higher than previous state-of-the-art global-scale ungauged basin tests. The baseline model without FDC was already competitive (median KGE 0.56), but integrating FDCs had substantial value. Because of the inaccurate representation of inputs, the baseline models might sometimes produce catastrophic results. However, model generalizability was further meaningfully improved by compiling an ensemble based on models with different input selections.

This paper presents new approach based on grammar induction called AMLSI Action Model Learning with State machine Interactions. The AMLSI approach does not require a training dataset of plan traces to work. AMLSI proceeds by trial and error: it queries the system to learn with randomly generated action sequences, and it observes the state transitions of the system, then AMLSI returns a PDDL domain corresponding to the system. A key issue for domain learning is the ability to plan with the learned domains. It often happens that a small learning error leads to a domain that is unusable for planning. Unlike other algorithms, we show that AMLSI is able to lift this lock by learning domains from partial and noisy observations with sufficient accuracy to allow planners to solve new problems.

The aim of this paper is to develop and test advanced analytical methods to improve the prediction accuracy of Credit Risk Models, preserving at the same time the model interpretability. In particular, the project focuses on applying an explainable machine learning model to PSD2-related databases. The input data were obtained solely from synthetic account transactions generated from a pool of commercial banks from a pool of Italian commercial banks. Over the total proven models, CatBoost has shown the highest performance. The algorithm implementation produces a GINI of 0.45 after tuning the hyper-parameters combined with their inherent class-weight resampling method. SHAP package is used to provide a global and local interpretation of the model predictions to formulate a human-comprehensive approach to understanding the decision-maker algorithm. The 20 most important features are selected using the Shapley values to present a full human-understandable model that reveals how the attributes of an individual are related to its model prediction.

Responsible artificial intelligence (AI) provides a framework for building trust in the AI solutions of an organisation, according to a report from Accenture. It is defined as the practice of designing, developing, and deploying AI with good intention to empower employees and businesses, and fairly impact customers and society. In turn, this allows companies to stimulate trust and scale AI with confidence. With technology starting to become commonplace, more organisations around the world are seeing the need to adopt responsible AI. For example, Microsoft relies on an AI, Ethics, and Effects in Engineering and Research (Aether) Committee to advise its leadership on the challenges and opportunities presented by AI innovations. Some of the elements the committee examines is how fairly AI systems treat people, the reliability and safety of AI systems, how AI systems empower employees to engage with one other, and how understandable AI systems are.