Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
A variational autoencoder-based order parameter is proposed and demonstrated on various high-entropy alloys, providing a computational technique for understanding chemical ordering in alloys. Ultimately, this has the potential to facilitate the development of rational alloy design strategies.
An algorithmic approach is developed to analyze large-scale patient safety data and remove the confounders of reporting trajectory and drug inference. Such an approach can be effectively used to investigate demographic disparities of drug safety and to identify at-risk patients during a pandemic.
The authors demonstrate how neural systems can encode cognitive functions, and use the proposed model to train robust, scalable deep neural networks that are explainable and capable of symbolic reasoning and domain generalization.
Combining human mobility data and nonlinear mathematical analysis techniques, this study offers insights into the interplay between human behavior, disease transmission and non-pharmaceutical interventions during the COVID-19 pandemic.
The authors propose Detect, a browser-based anomaly detection framework for diffusion magnetic resonance imaging tractometry data. The tool leverages normative microstructural brain features derived from healthy participants using deep autoencoders to detect anomalies at the individual level.
The authors propose EPICS, a method to predict microbial community structures by estimating effective pairwise interactions that subsume high-order interactions between species. EPICS is more efficient and applicable to larger communities than current approaches.
An efficient method for parallelizing the contraction of tensor networks pushes the boundaries for the classical simulation of quantum computation, and aids the development of quantum algorithms and hardware.
The study introduces the design and implementation of a parallel computational framework, called HiCOPS, for efficient acceleration of large-scale database peptide search workloads on supercomputers.
Optical computing promises high-speed computations but presents challenges in nonlinear information processing. This Article demonstrates a scalable and energy-efficient nonlinear optical-computing framework that can perform machine learning tasks.
The study develops a machine learning approach for predicting bone regeneration in an additively manufactured bioceramic scaffold, which is correlated with an in vivo sheep model, exhibiting effectiveness for solving such a multiscale modeling problem.
Countries are using hospital admission policies that prioritize patients with COVID-19 during the pandemic. The authors propose an alternative open-source framework to optimally schedule hospital care for all diseases and patients that can save life years overall.
This work demonstrates that large gains still exist in accelerating and improving the coverage of reaction prediction algorithms. These advances create opportunities for computationally exploring deeper and broader reaction networks.
The authors propose a deep learning model that analyzes single-cell RNA sequencing (scRNA-seq) data by explicitly modeling gene regulatory networks (GRNs), outperforming the state-of-art methods on various tasks, including GRN inference, scRNA-seq analysis and simulation.
An evidence-based approach for dealing with insufficient, conflicting and biased materials data is proposed for recommending high-entropy alloys, showing good capabilities for extrapolating the number of components.
Deep graph neural networks can refine a predicted protein model efficiently with less computing resources. The accuracy is comparable to that of the leading physics-based methods that rely on time-consuming conformation sampling.
Near-term quantum computers hold many promises but remain limited to a moderate number of qubits. This Article presents a pathway for modeling correlated materials with a reduced number of qubits, bringing quantum computing to materials modeling.
A class of quantum neural networks is presented that outperforms comparable classical feedforward networks. They achieve a higher capacity in terms of effective dimension and at the same time train faster, suggesting a quantum advantage.
BaseQTL is a Bayesian method to map molecular QTL affecting allele-specific expression even when no genotypes are available. It is well suited to discover eQTLs hidden in a wealth of RNA-seq data to unravel molecular mechanisms underpinning disease.
A statistical modeling method is proposed to generalize right censored data to a standard regression problem, thus making it possible to apply regression learning algorithms to survival prediction problems.
Predicting binding specificity of T-cell receptors (TCRs) and putative antigens can help improve cancer immunotherapy. Lin et al. propose RACER, which efficiently makes use of supervised machine learning to learn important molecular interactions contributing to TCR–peptide binding.