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.

Two threads of research in the quest for methods that predict the 3D structures of proteins from their amino-acid sequences have become fully intertwined. The result is a leap forward in the accuracy of predictions.

A microscopy system that enables simultaneous recording from hundreds of neurons in the mouse visual cortex reveals that the brain enhances its coding capacity by representing visual inputs in dimensions perpendicular to correlated noise.

Despite being only one-atom thick, defect-free graphene is considered to be completely impermeable to all gases and liquids(1-10). This conclusion is based on theory(3-8) and supported by experiments(1,9,10) that could not detect gas permeation through micrometre-size membranes within a detection limit of 10(5) to 10(6) atoms per second. Here, using small monocrystalline containers tightly sealed with graphene, we show that defect-free graphene is impermeable with an accuracy of eight to nine orders of magnitude higher than in the previous experiments. We are capable of discerning (but did not observe) permeation of just a few helium atoms per hour, and this detection limit is also valid for all other gases tested (neon, nitrogen, oxygen, argon, krypton and xenon), except for hydrogen. Hydrogen shows noticeable permeation, even though its molecule is larger than helium and should experience a higher energy barrier. This puzzling observation is attributed to a two-stage process that involves dissociation of molecular hydrogen at catalytically active graphene ripples, followed by adsorbed atoms flipping to the other side of the graphene sheet with a relatively low activation energy of about 1.0 electronvolt, a value close to that previously reported for proton transport(11,12). Our work provides a key reference for the impermeability of two-dimensional materials and is important from a fundamental perspective and for their potential applications.

Precision spectroscopy of atomic systems(1) is an invaluable tool for the study of fundamental interactions and symmetries(2). Recently, highly charged ions have been proposed to enable sensitive tests of physics beyond the standard model(2-5) and the realization of high-accuracy atomic clocks(3,5), owing to their high sensitivity to fundamental physics and insensitivity to external perturbations, which result from the high binding energies of their outer electrons. However, the implementation of these ideas has been hindered by the low spectroscopic accuracies (of the order of parts per million) achieved so far(6-8). Here we cool trapped, highly charged argon ions to the lowest temperature reported so far, and study them using coherent laser spectroscopy, achieving an increase in precision of eight orders of magnitude. We use quantum logic spectroscopy(9,10) to probe the forbidden optical transition in Ar-40(13+) at a wavelength of 441 nanometres and measure its excited-state lifetime and g-factor. Our work unlocks the potential of highly charged ions as ubiquitous atomic systems for use in quantum information processing, as frequency standards and in highly sensitive tests of fundamental physics, such as searches for dark-matter candidates(11) or violations of fundamental symmetries(2).

The precision of laser spectroscopy of highly charged ions is improved by eight orders of magnitude by cooling trapped, highly charged ions and using quantum logic spectroscopy, thereby enabling tests of fundamental physics.