In the summer of 1956, a small group of mathematicians and computer scientists gathered at Dartmouth College to embark on the grand project of designing intelligent machines. The ultimate goal, as they saw it, was to build machines rivaling human intelligence. As the decades passed and AI became an established field, it lowered its sights. There were great successes in logic, reasoning, and game-playing, but stubborn progress in areas like vision and fine motor-control. This led many AI researchers to abandon their earlier goals of fully general intelligence, and focus instead on solving specific problems with specialized methods.

The last decade has seen a vast increase both in the diversity of applications to which machine learning is applied, and to the import of those applications. Machine learning is no longer just the engine behind ad placements and spam filters; it is now used to filter loan applicants, deploy police officers, and inform bail and parole decisions, among other things. The result has been a major concern for the potential for data-driven methods to introduce and perpetuate discriminatory practices, and to otherwise be unfair. And this concern has not been without reason: a steady stream of empirical findings has shown that data-driven methods can unintentionally both encode existing human biases and introduce new ones.7,9,11,60 At the same time, the last two years have seen an unprecedented explosion in interest from the academic community in studying fairness and machine learning. "Fairness and transparency" transformed from a niche topic with a trickle of papers produced every year (at least since the work of Pedresh56 to a major subfield of machine learning, complete with a dedicated archival conference--ACM FAT*). But despite the volume and velocity of published work, our understanding of the fundamental questions related to fairness and machine learning remain in its infancy.

Women are underrepresented in the fields of science, technology, engineering, and mathematics (STEM) in most countries, including Germany and the U.S.29,32 This was demonstrated in several surveys investigating the proportion of women in the STEM fields for specific populations. Some of these studies, for example, investigated the number of enrolled students10,30 or the percentage of female professors at universities. Other studies analyzed the disparities in research funding.23 Nearly all these surveys selected a particular population of women in consideration of their university degree or their nationality.11,34 Like many other studies investigating the gender gap and its reasons in science, these surveys are usually based on data records from several kinds of registrations or enrollments, for example, the enrollment as student or doctoral student, the registration of finished doctoral theses or the membership as professor in a certain country.1,14,16,28 However, researchers at the postdoctoral level or industrial researchers are often not registered and unfortunately drop out of the surveys.

The spread of the novel coronavirus that causes COVID-19 has prompted state and local governments around the U.S. to institute shelter-in-place orders and business closures. As millions suddenly find themselves confined to their homes, the shift has strained not only internet service providers, streaming platforms, and online retailers, but the utilities supplying power to the nation's electrical grid, as well. U.S. electricity use on March 27, 2020 was 3% lower than it was on March 27, 2019, a loss of about three years of sales growth. Peter Fox-Penner, director of the Boston University Institute for Sustainable Energy, asserted in a recent op-ed that utility revenues will suffer because providers are halting shutoffs and deferring rate increases. Moreover, according to research firm Wood Mackenzie, the rise in household electricity demand won't offset reduced business electricity demand, mainly because residential demand makes up just 40% of the total demand across North America.

The NHS is turning to artificial intelligence (AI) to help predict upcoming demand for intensive care beds and ventilators during the coronavirus pandemic across England. Trials of the predictive system, known as the COVID 19 Capacity Planning and Analysis System (CPAS), began today at four hospitals. It harnesses the principles of machine learning – algorithms that find and apply patterns in data – to provide statistics, forecasts and simulation environments to the NHS to better plan resources during the pandemic. For example, predictions made by the machine learning system could inform a hospital that capacity will be reached in advance, giving it time to bring in extra resources or share capacity with neighbouring hospitals. If CPAS proves to be accurate, the NHS will look to roll it out across the rest of the country.

Machine learning is leading to numerous changes in the energy industry. The Department of Energy recently announced that it is taking steps to accelerate the integration of machine learning technology in energy research and development. The head of the Department of Energy announced that they will be investing $30 million in artificial intelligence and machine learning algorithms. The new programs will have multiple purposes. One of the biggest goals is to use machine learning to facilitate the development of new renewable energy technologies.

Companies using artificial intelligence (AI) across their business units should consider creating a C-suite position to oversee how AI is used and guard against the risk of making bad decisions based on biased algorithms, experts say. Only a few companies, like Levi Strauss & Co, have established a chief artificial intelligence officer (CAIO) position, and fewer have created a C-level position dedicated solely to AI ethics. Brian Kropp, chief of research in the HR practice at Gartner, said chief technology officers and chief information officers will struggle with handling AI-related decisions and ethical dilemmas. "CTOs and CIOs are going to be thinking about the role through the lens of how they can make the technology work," Kropp said. However, "artificial intelligence is not a question of how you get the technology to work; it's a question of how do you think through the implications of the technology?"

Researchers demonstrated the mathematical impossibility of representing social networks and other complex networks using popular methods of'low-dimensional embeddings.' Models and algorithms for analyzing complex networks are widely used in research and affect society at large through their applications in online social networks, search engines, and recommender systems. According to a new study, however, one widely used algorithmic approach for modeling these networks is fundamentally flawed, failing to capture important properties of real-world complex networks. "It's not that these techniques are giving you absolute garbage. They probably have some information in them, but not as much information as many people believe," said C. "Sesh" Seshadhri, associate professor of computer science and engineering in the Baskin School of Engineering at UC Santa Cruz. Seshadhri is first author of a paper on the new findings published on March 2, 2020, in Proceedings of the National Academy of Sciences.

In a new study published this week on the preprint server Arxiv.org, Both it and several baseline models will soon be available on GitHub. One of the most powerful techniques in machine learning -- reinforcement learning, which entails spurring software agents toward goals via rewards -- is also one of the most flawed. It's sample inefficient, meaning it requires a large number of compute cycles to complete, and without additional data to cover variations, it adapts poorly to environments that differ from the training environment. It's theorized that prior knowledge of tasks through structured language could be combined with reinforcement learning to mitigate its shortcomings, and BabyAI was designed to put this theory to the test.

Every day for the past few weeks, charts and graphs plotting the projected apex of COVID-19 infections have been splashed across newspapers and cable news. Many of these models have been built using data from studies on previous outbreaks like SARS or MERS. Now, a team of engineers at MIT has developed a model that uses data from the COVID-19 pandemic in conjunction with a neural network to determine the efficacy of quarantine measures and better predict the spread of the virus. "Our model is the first which uses data from the coronavirus itself and integrates two fields: machine learning and standard epidemiology," explains Raj Dandekar, a Ph.D. candidate studying civil and environmental engineering. Together with George Barbastathis, professor of mechanical engineering, Dandekar has spent the past few months developing the model as part of the final project in class 2.168 (Learning Machines).