Effective conservation actions require effective population monitoring. However, accurately counting animals in the wild to inform conservation decision-making is difficult. Monitoring populations through image sampling has made data collection cheaper, wide-reaching and less intrusive but created a need to process and analyse this data efficiently. Counting animals from such data is challenging, particularly when densely packed in noisy images. Attempting this manually is slow and expensive, while traditional computer vision methods are limited in their generalisability. Deep learning is the state-of-the-art method for many computer vision tasks, but it has yet to be properly explored to count animals. To this end, we employ deep learning, with a density-based regression approach, to count fish in low-resolution sonar images. We introduce a large dataset of sonar videos, deployed to record wild mullet schools (Mugil liza), with a subset of 500 labelled images. We utilise abundant unlabelled data in a self-supervised task to improve the supervised counting task. For the first time in this context, by introducing uncertainty quantification, we improve model training and provide an accompanying measure of prediction uncertainty for more informed biological decision-making. Finally, we demonstrate the generalisability of our proposed counting framework through testing it on a recent benchmark dataset of high-resolution annotated underwater images from varying habitats (DeepFish). From experiments on both contrasting datasets, we demonstrate our network outperforms the few other deep learning models implemented for solving this task. By providing an open-source framework along with training data, our study puts forth an efficient deep learning template for crowd counting aquatic animals thereby contributing effective methods to assess natural populations from the ever-increasing visual data.

In this paper, we propose a parametrised factor that enables inference on Gaussian networks where linear dependencies exist among the random variables. Our factor representation is a generalisation of traditional Gaussian parametrisations where the positive-definite constraint (of covariance and precision matrices) has been relaxed. For this purpose, we derive various statistical operations and results (such as marginalisation, multiplication and affine transformations of random variables) which extend the capabilities of Gaussian factors to these degenerate settings. By using this principled factor definition, degeneracies can be accommodated accurately and automatically at little additional computational cost. As illustration, we apply our methodology to a representative example involving recursive state estimation of cooperative mobile robots.

Most graph neural network architectures work by message-passing node vector embeddings over the adjacency matrix, and it is assumed that they capture graph topology by doing that. We design two synthetic tasks, focusing purely on topological problems - triangle detection and clique distance - on which graph neural networks perform surprisingly badly, failing to detect those "bermuda" triangles. Many tasks need to handle the graph representation of data in areas such as chemistry (Wale & Karypis, Method Triangles Clique 2006), social networks (Fan et al., 2019), and transportation GCN 50.0 50.0 (Zhao et al., 2019). Furthermore, it is not GCN D 75.7 83.2 limited to these graph tasks but also includes images GCN D ID 80.4 83.4 (Chen et al., 2019) and 3D polygons (Shi & Rajkumar, GIN 74.1 97 2020) that are possible to convert to graph data GIN D 75.0 99.4 formats. Because of these broad applications, Graph GIN D ID 70.5 100.0 Deep Learning is an important field in machine learning GAT 50.0 50.0 research. GAT D 88.5 99.9 Graph neural networks (GNNs, (Scarselli et al., 2008)) GAT D ID 94.1 100.0 is a common approach to perform machine learning SVM WL 67.2 73.1 with graphs. Most graph neural networks update SVM Graphlets 99.6 60.3 the graph node vector embeddings using the message passing. Node vector embeddings are usually initialized FCNN 55.6 54.6 with data features and local graph features like TF 100.0 70.0 node degrees. Then, for a (n 1)-th stacked layer, the TF AM 100.0 100.0 new node state is computed from the node vector representation TF-IS AM 86.7 100.0 of the previous layer (n).

Artificial Intelligence (AI) can augment and sometimes even replace human cognition. Inspired by efforts to value human agency alongside productivity, we discuss the benefits of solving Citizen Science (CS) tasks with Hybrid Intelligence (HI), a synergetic mixture of human and artificial intelligence. Currently there is no clear framework or methodology on how to create such an effective mixture. Due to the unique participant-centered set of values and the abundance of tasks drawing upon both human common sense and complex 21st century skills, we believe that the field of CS offers an invaluable testbed for the development of HI and human-centered AI of the 21st century, while benefiting CS as well. In order to investigate this potential, we first relate CS to adjacent computational disciplines. Then, we demonstrate that CS projects can be grouped according to their potential for HI-enhancement by examining two key dimensions: the level of digitization and the amount of knowledge or experience required for participation. Finally, we propose a framework for types of human-AI interaction in CS based on established criteria of HI. This "HI lens" provides the CS community with an overview of several ways to utilize the combination of AI and human intelligence in their projects. It also allows the AI community to gain ideas on how developing AI in CS projects can further their own field.

Have you interacted with a digital persona yet? At the Museum of Art & Photography in Bangalore, you can have a deep and engaging exchange with one that represents the late artist M.F. Husain -- considered the "Picasso of India" by many. This avatar is eager to talk art. And if you ask him whether he's real, he will look straight at you and say, "As close to real, enough to impress you."

SCI COMMUN### Climate change The United States submitted its new goals to the Paris climate agreement last week, pledging to cut its planet-warming greenhouse gas emissions 50% to 52% below 2005 levels by 2030. The nonbinding pledge, made on Earth Day at an online climate summit convened by President Joe Biden, is one of the most aggressive targets of any wealthy country. The cuts are also higher than the 26% to 28% reduction by 2025 pledged by former President Barack Obama when the United States first joined the agreement in 2015. In the past month, Canada, Japan, and the United Kingdom have also committed to cuts steeper than their earlier pledges; other large polluters, such as China, India, and Russia, have yet to increase their goals ahead of a critical U.N. climate meeting this winter in Glasgow, U.K. The U.S. pledge has a long road to reality. It will require immediate increases in renewable energy, widespread adoption of electric vehicles, and other steps, many of which will require laws that could be difficult to pass, such as Biden's climate-focused infrastructure bill. > “It is our policy … not to employ anyone who has taken the experimental COVID-19 injection.” > > Miami-based private school Centner Academy in a letter to parents this week, citing discredited claims that vaccinated people can transmit harmful substances to others. ### Biomedicine ![Figure][1] CREDITS: (GRAPHIC) J. BRAINARD/ SCIENCE ; (DATA) ASGCT/INFORMA The number of new clinical trials for gene, cell, and RNA therapies has almost tripled in the past 4 years, the American Society of Gene & Cell Therapy said this month in its first ever quarterly trends report. As of 31 March, nearly 3500 of these experimental treatments were in development, most of which (53%) consist of altered genes or genetically modified (GM) cells, such as cancer-fighting T cells with modified receptors. Cancer is the target for the largest number (1200) of therapies under development. Among RNA therapies, many were vaccines (35) or COVID-19 treatments (30). The United States has more clinical trials underway in each of the three types of therapies than any other country: 1400 overall. Globally, 16 gene therapies (including GM cells), 53 non-GM cell therapies, and 15 RNA therapies have been approved for use so far. ### Vaccination The Department of Health and Human Services (HHS) last week killed a rule that would have made it much more difficult for people who sustain shoulder injuries during vaccination to win compensation from a $4.1 billion government fund. The rule had been finalized on 19 January, the last day of former President Donald Trump's administration. But President Joe Biden's administration froze its implementation on 20 January. The new administration said the previous one had been “irregular in its haste” when it moved to remove shoulder injuries from a list of injuries in which the petitioner does not have to prove a vaccine caused the injury, making it easier to win a government payout ( Science , 10 April 2020, p. [121][2]). Shoulder injuries accounted for nearly 55% of more than 2400 claims filed with the National Vaccine Injury Compensation Program in the past 2 years, most of them after flu shots. COVID-19 vaccines fall under a different HHS program, but if they win formal approval by the U.S. Food and Drug Administration, they could in principle be added to the national compensation program. ### Publishing Sixteen journals, including BMJ Open Science and Royal Society Open Science , say they will accept articles reviewed by the nonprofit Peer Community In Registered Reports (PCI RR). The organization, launched last week, will review one type of article: “registered reports,” describing studies for which detailed experimental plans are peer reviewed before research begins. Once the research is complete, PCI RR will do a second round of peer review including results and analysis. Papers it recommends can then be published in any of the 16 journals without further review as long as they meet a journal's normal criteria. The organization, funded by donations, will provide peer review free to authors and journals in any discipline. ### COVID-19 The National Institutes of Health (NIH) last week announced it would launch a large study repurposing existing drugs for patients with mild COVID-19 symptoms who don't need hospitalization. The $155 million trial, which aims to recruit 13,500 participants, will open in a few weeks at several research centers. It will test up to seven medications already approved for other conditions, but NIH hasn't yet named them. A recent, similar trial at the University of Oxford that has so far enrolled about 4800 patients has found that the asthma drug budesonide lessens symptoms and speeds recovery in certain COVID-19 patients. ### Medicine A large study has firmed up earlier evidence that SARS-CoV-2 increases the rate of complications for pregnant women and their babies. The study followed 706 pregnant women with COVID-19 and 1424 uninfected pregnant women at hospitals in 18 countries. Infected women had a 76% higher risk of developing problems caused by pregnancy-associated high blood pressure and a 59% higher risk of preterm birth. They were also five times more likely to be admitted to intensive care than uninfected women. Eleven women with COVID-19 died, compared with one uninfected woman, researchers report in JAMA Pediatrics . Infected women with fever and shortness of breath had babies with a fivefold increased risk of complications such as immature lungs, eye disorders, and brain damage. The coronavirus may affect pregnancy via changes in a woman's heart, lungs, and immune system. The results show pregnant women should be among priority groups for COVID-19 vaccines, the authors say. A separate study last week found no obvious safety problems in more than 800 U.S. women who gave birth after receiving messenger RNA vaccines. ### Community The American Humanist Association has decided to withdraw Richard Dawkins's 1996 Humanist of the Year award as a result of his “history of making statements that use the guise of scientific discourse to demean marginalized groups.” The decision came soon after the evolutionary biologist and former University of Oxford professor wrote a tweet comparing transgender people to Rachel Dolezal, a civil rights activist who for years posed as Black. The association said the tweet “implies that the identities of transgender individuals are fraudulent, while also simultaneously attacking Black identity as one that can be assumed when convenient.” In 2015, Dawkins argued that trans women are not women based on their chromosomes, but said he would use the pronoun “she” out of courtesy. ### Seismology The first large-scale, phone-based earthquake early warning system will be deployed in Greece and New Zealand, Google announced this week. Since last year, the company has been testing the use of data compiled from its more than 2 billion active Android phones to pinpoint the location and strength of earthquakes. The measurement comes from the built-in Android phones' accelerometers, which sense movement just like seismometers. When the phones detect earthquakelike signals, they alert a server that combines information from many phones. If enough phones corroborate the result, an alert goes out. The cellphone results compared well to those from seismometer-based warning systems in Japan and the United States. Google chose New Zealand and Greece, both countries with high earthquake hazards and many Android phones, to premier the system because they lack operational warning systems of their own. Eventually, phone-based alerts could be available worldwide. ### Biotechnology Genetically modified mosquitoes designed to prevent the spread of viruses such as Zika and dengue are set to be released in the United States for the first time. Starting this week, the company Oxitec will free fewer than 12,000 transgenic Aedes aegypti mosquitoes, all of them nonbiting males, in the Florida Keys as part of a pilot study, the company announced on 23 April. They are engineered to carry a gene that kills their female offspring, reducing the population of mosquitoes capable of transmitting diseases. Field tests outside the United States have shown dramatic population drops, though the company has not published definitive evidence that the strategy reduces disease in humans. The project, which has long faced public opposition in Florida, won approval in May 2020 from the Environmental Protection Agency, which predicted no adverse effects on people or local wildlife. ### Science policy President Joe Biden last week picked two veterans of government service and a newcomer to fill top science positions. He named soil scientist Asmeret Asefaw Berhe of the University of California, Merced, to lead the Department of Energy's Office of Science. Berhe, born in Eritrea, has little government experience but has won accolades for her research and efforts to promote diversity in science. She would become the first Black woman to lead the science office if confirmed by the Senate. Oceanographer Rick Spinrad of Oregon State University, Corvallis, who has held numerous posts at the National Oceanic and Atmospheric Administration, is Biden's choice to lead that agency. To run the State Department's science bureau, he chose Monica Medina, an ocean policy expert and attorney at Georgetown University. Confirmation hearings for the three nominees could come as early as next month. ### Regulation In one of the broadest attempts to regulate artificial intelligence (AI) to date, the European Commission on 21 April proposed new rules for algorithms that power everything from medical device and credit scoring software to chatbots and facial recognition systems. The rules divide AI technologies into risk categories, and put outright bans on some, such as systems that would score individuals' “social credit.” Other “high-risk” systems, including those that collect biometric data, would require a strict vetting process. However, there are exceptions for national security, and it could take years before the rules become law: They must first pass the European Council and the European Parliament and be adopted by member countries. ### Community An effort to increase gender, racial, and geographic diversity in the U.S. National Academy of Sciences (NAS) has begun to bear fruit. Nearly half of the 120-member 2021 class announced last week are women, compared with one-quarter in 2011. The new cohort includes nine Black scientists; NAS officials say previous classes never had more than three and often had none. “We need to do better, but I'm amazed at how far we've come,” says plant geneticist Susan Wessler, NAS home secretary. The academy's governing council has begun to give more slots to disciplinary units that present candidate slates less skewed toward older white men. To reduce its geographic imbalance, NAS is also prohibiting members from nominating someone from their own institution. Today, 18 U.S. states have two or fewer members, whereas a handful of elite academic institutions each have more than 100. [1]: pending:yes [2]: http://www.sciencemag.org/content/368/6487/121

Opened in November 2020, Hotel Sky in Sandton, Johannesburg, made its debut with three robots: Lexi, Micah, and Ariel. Lending a helpful hand to the human staff at the property, these robots are the hotel's answer to travelers' increased desire for socially distant interactions. Lexi, Micah, and Ariel can deliver room service, provide travel information, and carry up to 165 pounds of luggage each from the marble-floored lobby to the rooms.

By using the power of Reveal Data's first class neural-net AI, along with the data and know-how available within a global law firm like DLA Piper, the AI has been taught to spot these cartel risks and enable our lawyers to quickly run a review and advise clients of their cartel risks. Specifically, Aiscension has been trained to uncover the following forms of cartel behavior: price fixing; bid rigging; market sharing; collective boycotts and exchanging competitively sensitive information.

The use of deep neural networks (DNNs) in safety-critical applications like mobile health and autonomous driving is challenging due to numerous model-inherent shortcomings. These shortcomings are diverse and range from a lack of generalization over insufficient interpretability to problems with malicious inputs. Cyber-physical systems employing DNNs are therefore likely to suffer from safety concerns. In recent years, a zoo of state-of-the-art techniques aiming to address these safety concerns has emerged. This work provides a structured and broad overview of them. We first identify categories of insufficiencies to then describe research activities aiming at their detection, quantification, or mitigation. Our paper addresses both machine learning experts and safety engineers: The former ones might profit from the broad range of machine learning topics covered and discussions on limitations of recent methods. The latter ones might gain insights into the specifics of modern ML methods. We moreover hope that our contribution fuels discussions on desiderata for ML systems and strategies on how to propel existing approaches accordingly.

The increasing polarization of online political discourse calls for computational tools that are able to automatically detect and monitor ideological divides in social media. Here, we introduce a minimally supervised method that directly leverages the network structure of online discussion forums, specifically Reddit, to detect polarized concepts. We model polarization along the dimensions of agenda setting and framing, drawing upon insights from moral psychology. The architecture we propose combines graph neural networks with structured sparsity and results in representations for concepts and subreddits that capture phenomena such as ideological radicalization and subreddit hijacking. We also create a new dataset of political discourse covering 12 years and more than 600 online groups with different ideologies.