Browse Articles

Predicting stunting status among under five children in ethiopia using ensemblemachine learning algorithms

Advancing deep learning for expressive music composition and performance modeling

The ocean's deepest animal ecosystem

A qualitative study on ethical issues related to the use of AI-driven technologies in foreign language learning

My AI chatbot thinks my idea is fundable

Identifying monthly rainfall erosivity patterns using hourly rainfall data across India

Abstract Rainfall erosivity is a key dynamic factor of water erosion estimation, with a significant spatial and temporal variation. This study presents a comprehensive analysis of the spatial patterns and monthly distribution of rainfall erosivity across India, using data from 261 hourly and 2,525 monthly rainfall stations covering the period from 1969 to 2021. In India, monthly rainfall erosivity and related attributes—such as the kinetic energy of erosive rainfall, the number of erosive events, and peak hourly rainfall intensity—have been systematically examined for the first time. Monthly erosivity estimates derived from hourly data were linked with monthly rainfall, enabling a simplified and efficient estimation approach. To predict monthly erosivity based on rainfall, temperature, and topographic variables, we developed and evaluated three modeling approaches: linear regression, a machine learning-based XGBoost model, and an ensemble model. XGBoost outperformed the others, achieving a median coefficient of determination (R2) of 0.97, while the ensemble model also performed well with a median R2 of 0.96. Additionally, a Geographically Weighted Regression (GWR) approach was applied for spatial interpolation, yielding accurate high-resolution erosivity maps with a median R2 of 0.90. The results also demonstrate that erosivity peaks during the summer monsoon months (June to September), with July exhibiting the highest value due to intense rainfall and high kinetic energy. Notably, the analysis revealed that nearly 32% of India experiences monthly erosivity exceeding 2,000 MJ mm ha−1  h−1  month−1  in July alone. In contrast, non-monsoon months showed considerably lower erosivity levels across most of the country. A statistically significant long-term increase was detected in January, with an average rise of +0.86 MJ mm ha−1  h−1  month−1  in total erosivity and + 0.1 mm h−1  in maximum 60-min rainfall intensity annually. While acknowledging certain limitations, this study provides valuable insights into erosive rainfall characteristics, enhances rain-driven erosion assessment, and supports the development of timely and location-specific soil conservation strategies across India.

Daily briefing: Respiratory illness can ‘wake up’ dormant cancer cells

The relationship between parental psychological well-being and preadolescents’ social media use

Genome-wide analysis of the RING-H2 E3 ubiquitin ligase SlATL family in tomato

Increased risk of atrial fibrillation in uterine fibroid patients: a nationwide population-based study

Correction: Gender differences in the association between community care and quality of life among older adults in China

Survey of US postdocs finds threefold increase in job losses

Fusion of SARS-CoV-2 neutralizing LCB1 peptide with Bacillus amyloliquefaciens RNase improves antiviral efficacy

Author Correction: Genome-wide CRISPR screen in human T cells reveals regulators of FOXP3

Elevated stress hyperglycemia ratio associated with higher hospital mortality in patients with respiratory failure

RpoB mutation patterns in Rifampicin-resistant tuberculosis: a Jiangxi Province study, 2021–2023

Direct observation of coherent elastic antineutrino–nucleus scattering

Abstract Neutrinos are elementary particles that interact only very weakly with matter. Neutrino experiments are, therefore, usually big, with masses in the multi-tonne range. The thresholdless interaction of coherent elastic scattering of neutrinos on atomic nuclei leads to greatly enhanced interaction rates, which allows for much smaller detectors. The study of this process gives insights into physics beyond the Standard Model of particle physics. The CONUS+ experiment1 was designed to first detect elastic neutrino–nucleus scattering in the fully coherent regime with low-energy neutrinos produced in nuclear reactors. For this purpose, semiconductor detectors based on high-purity germanium crystals with extremely low-energy thresholds were developed2. Here we report the first observation of a neutrino signal with a statistical significance of 3.7σ from the CONUS+ experiment, operated at the nuclear power plant in Leibstadt, Switzerland. In 119 days of reactor operation (395 ± 106) neutrinos were measured compared with a predicted number from calculations assuming Standard Model physics of (347 ± 59) events. With increased precision, there is potential for fundamental discoveries in the future. The CONUS+ results in combination with other measurements of this interaction channel might therefore mark a starting point for a new era in neutrino physics.

A human-specific enhancer fine-tunes radial glia potency and corticogenesis

Novel anti-VEGF scFv antibodies with superior in vitro and in vivo activities

Whole-body physics simulation of fruit fly locomotion

Abstract The body of an animal influences how its nervous system generates behaviour1. Accurately modelling the neural control of sensorimotor behaviour requires an anatomically detailed biomechanical representation of the body. Here we introduce a whole-body model of the fruit fly Drosophila melanogaster in a physics simulator2. Designed as a general-purpose framework, our model enables the simulation of diverse fly behaviours, including both terrestrial and aerial locomotion. We validate its versatility by replicating realistic walking and flight behaviours. To support these behaviours, we develop phenomenological models for fluid and adhesion forces. Using data-driven, end-to-end reinforcement learning3,4, we train neural network controllers capable of generating naturalistic locomotion5–7 along complex trajectories in response to high-level steering commands. Furthermore, we show the use of visual sensors and hierarchical motor control8, training a high-level controller to reuse a pretrained low-level flight controller to perform visually guided flight tasks. Our model serves as an open-source platform for studying the neural control of sensorimotor behaviour in an embodied context.