资源环境科技发展态势分析平台

Screening mammography aims to identify breast cancer at earlier stages of the disease, when treatment can be more successful(1). Despite the existence of screening programmes worldwide, the interpretation of mammograms is affected by high rates of false positives and false negatives(2). Here we present an artificial intelligence (AI) system that is capable of surpassing human experts in breast cancer prediction. To assess its performance in the clinical setting, we curated a large representative dataset from the UK and a large enriched dataset from the USA. We show an absolute reduction of 5.7% and 1.2% (USA and UK) in false positives and 9.4% and 2.7% in false negatives. We provide evidence of the ability of the system to generalize from the UK to the USA. In an independent study of six radiologists, the AI system outperformed all of the human readers: the area under the receiver operating characteristic curve (AUC-ROC) for the AI system was greater than the AUC-ROC for the average radiologist by an absolute margin of 11.5%. We ran a simulation in which the AI system participated in the double-reading process that is used in the UK, and found that the AI system maintained non-inferior performance and reduced the workload of the second reader by 88%. This robust assessment of the AI system paves the way for clinical trials to improve the accuracy and efficiency of breast cancer screening.

Many corals harbour symbiotic dinoflagellate algae. The algae live inside coral cells in a specialized membrane compartment known as the symbiosome, which shares the photosynthetically fixed carbon with coral host cells while host cells provide inorganic carbon to the algae for photosynthesis(1). This endosymbiosis-which is critical for the maintenance of coral reef ecosystems-is increasingly threatened by environmental stressors that lead to coral bleaching (that is, the disruption of endosymbiosis), which in turn leads to coral death and the degradation of marine ecosystems(2). The molecular pathways that orchestrate the recognition, uptake and maintenance of algae in coral cells remain poorly understood. Here we report the chromosome-level genome assembly of aXeniaspecies of fast-growing soft coral(3), and use this species as a model to investigate coral-alga endosymbiosis. Single-cell RNA sequencing identified 16 cell clusters, including gastrodermal cells and cnidocytes, inXeniasp. We identified the endosymbiotic cell type, which expresses a distinct set of genes that are implicated in the recognition, phagocytosis and/or endocytosis, and maintenance of algae, as well as in the immune modulation of host coral cells. By couplingXeniasp. regeneration and single-cell RNA sequencing, we observed a dynamic lineage progression of the endosymbiotic cells. The conserved genes associated with endosymbiosis that are reported here may help to reveal common principles by which different corals take up or lose their endosymbionts.

Data analysis workflows in many scientific domains have become increasingly complex and flexible. Here we assess the effect of this flexibility on the results of functional magnetic resonance imaging by asking 70 independent teams to analyse the same dataset, testing the same 9 ex-ante hypotheses(1). The flexibility of analytical approaches is exemplified by the fact that no two teams chose identical workflows to analyse the data. This flexibility resulted in sizeable variation in the results of hypothesis tests, even for teams whose statistical maps were highly correlated at intermediate stages of the analysis pipeline. Variation in reported results was related to several aspects of analysis methodology. Notably, a meta-analytical approach that aggregated information across teams yielded a significant consensus in activated regions. Furthermore, prediction markets of researchers in the field revealed an overestimation of the likelihood of significant findings, even by researchers with direct knowledge of the dataset(2-5). Our findings show that analytical flexibility can have substantial effects on scientific conclusions, and identify factors that may be related to variability in the analysis of functional magnetic resonance imaging. The results emphasize the importance of validating and sharing complex analysis workflows, and demonstrate the need for performing and reporting multiple analyses of the same data. Potential approaches that could be used to mitigate issues related to analytical variability are discussed.

The results obtained by seventy different teams analysing the same functional magnetic resonance imaging dataset show substantial variation, highlighting the influence of analytical choices and the importance of sharing workflows publicly and performing multiple analyses.

The measurement sensitivity of quantum probes using N uncorrelated particles is restricted by the standard quantum limit(1), which is proportional to 1/root N. This limit, however, can be overcome by exploiting quantum entangled states, such as spin-squeezed states(2). Here we report the measurement-based generation of a quantum state that exceeds the standard quantum limit for probing the collective spin of 10(11) rubidium atoms contained in a macroscopic vapour cell. The state is prepared and verified by sequences of stroboscopic quantum non-demolition (QND) measurements. We then apply the theory of past quantum states(3,4) to obtain spin state information from the outcomes of both earlier and later QND measurements. Rather than establishing a physically squeezed state in the laboratory, the past quantum state represents the combined system information from these prediction and retrodiction measurements. This information is equivalent to a noise reduction of 5.6 decibels and a metrologically relevant squeezing of 4.5 decibels relative to the coherent spin state. The past quantum state yields tighter constraints on the spin component than those obtained by conventional QND measurements. Our measurement uses 1,000 times more atoms than previous squeezing experiments(5-10), with a corresponding angular variance of the squeezed collective spin of 4.6 x 10(-13) radians squared. Although this work is rooted in the foundational theory of quantum measurements, it may find practical use in quantum metrology and quantum parameter estimation, as we demonstrate by applying our protocol to quantum enhanced atomic magnetometry.

Since its introduction, the reward prediction error theory of dopamine has explained a wealth of empirical phenomena, providing a unifying framework for understanding the representation of reward and value in the brain(1-3). According to the now canonical theory, reward predictions are represented as a single scalar quantity, which supports learning about the expectation, or mean, of stochastic outcomes. Here we propose an account of dopamine-based reinforcement learning inspired by recent artificial intelligence research on distributional reinforcement learning(4-6). We hypothesized that the brain represents possible future rewards not as a single mean, but instead as a probability distribution, effectively representing multiple future outcomes simultaneously and in parallel. This idea implies a set of empirical predictions, which we tested using single-unit recordings from mouse ventral tegmental area. Our findings provide strong evidence for a neural realization of distributional reinforcement learning.

Analyses of single-cell recordings from mouse ventral tegmental area are consistent with a model of reinforcement learning in which the brain represents possible future rewards not as a single mean of stochastic outcomes, as in the canonical model, but instead as a probability distribution.