Science

All Biology & Medicine Physics & Materials Chemistry Astronomy & Space Mathematics Climate

98 stories · page 1 of 7

Evidence-backed gain

Developing and evaluating automated deep learning and human-in-the-loop vision-language systems for microplastic characterization

Microplastic (MP) pollution poses escalating environmental risks, demanding efficient and reproducible tools for morphological characterization of plastic particles. Traditional manual microscopy is labour-intensive, operator-dependent, and poorly suited to large-scale monitoring. This study presents a comparative evaluation of two distinct artificial intelligence paradigms for the analysis of optical microscope images of microplastics. The first paradigm is a domain-specific, multi-task deep learning (DL) class…

Scientific Reports · Climate

Evidence-backed problem

US building twice as much gas-fired capacity as China in AI boom, analysis finds

The US has surged ahead of China in the building of new gas-fired power generation, largely to feed a boom in artificial intelligence (AI) that is adding vast amounts of planet-heating emissions, a new analysis has found. For decades, China’s rapid economic growth has seen it outpace the US in the addition of new gas power generation but a recent “frenzy” in datacenter construction for AI has reversed this, said Global Energy Monitor (Gem) in its new report. The US is now building twice as much gas-fired capacit…

Climate

Both readings

AI-Assisted Thematic Analysis in a Quality Improvement Evaluation: Replicable Findings and Subtle Misrepresentations

Using artificial intelligence to identify themes in interview data from a quality improvement program evaluation produced 4 replicable themes grounded in the data. However, 2 consistently identified themes resulted from subtle misrepresentations and could have easily misled results without thorough data knowledge and output audit by the human research team.

Science

Evidence-backed gain

General Atomics and SDSC use AI to advance fusion laser research

Collaboration combines high-performance computing and artificial intelligence to accelerate the design and optimisation of laser systems for inertial confinement fusion research

Science

Both readings

Impact of allometric reference selection on management scale aboveground biomass estimation via Sentinel-2 and machine learning

Accurate estimation of aboveground biomass (AGB) is essential for sustainable forest management, carbon accounting, and climate change mitigation. In remote sensing-based biomass mapping, field-derived AGB values are commonly used as reference data; however, these values are strongly influenced by the selected allometric equation. This study evaluates how alternative allometric reference datasets affect Sentinel-2-based AGB estimation at the forest management scale in Pinus brutia stands. Reference AGB values we…

Climate

Evidence-backed problem

Evaluating machine learning and neural network architectures for forensic sex estimation using mandibular ramus and notch features on panoramic radiographs

Estimating a biological profile, such as sex, is a fundamental step in forensic identification when primary identifiers are unavailable for direct individual comparison. In forensic scenarios involving advanced decay, specific taphonomic alterations, or midfacial blunt force impacts, the mandibular ramus serves as a valuable anatomical marker due to its distinct sexual dimorphism and thick cortical structure, making it more resilient to fragmentation than other, more fragile facial bones. Despite its utility, th…

Science

Evidence-backed gain

Optimal Initialization Scale for Neural Networks With Locally Quadratic Loss Landscapes: An SGD Dynamics Perspective

Stochastic gradient descent (SGD), one of the most fundamental optimization algorithms in machine learning (ML), can be recast through a continuous-time approximation as a Fokker-Planck equation for Langevin dynamics, a viewpoint that has motivated many theoretical studies. Within this framework, we study the relationship between the quasi-stationary distribution derived from this equation and the initial distribution through the Kullback-Leibler (KL) divergence. As the quasi-steady-state distribution depends on…

Science

Productivity-based sampling for field-scale soil organic carbon mapping: from regional models to carbon farming needs

This study evaluates an innovative field-scale targeted sampling strategy within a regional hybrid spatial prediction model that combines machine learning and geostatistics. The framework is designed so that newly collected field observations are incorporated only through the local residual kriging step, while the regional Random Forest trend model remains unchanged, allowing field-scale predictions to be refined without full model refitting. The proposed sampling approach integrates a Normalized Difference Vege…

Climate

Productivity-based sampling for field-scale soil organic carbon mapping: from regional models to carbon farming needs

Quantum-Enhanced Transfer Learning for IDR Binding Partner Prediction

Intrinsically Disordered Regions (IDRs) play essential roles in cellular processes through interactions with proteins, nucleic acids, lipids, and metal ions, yet predicting their binding partners remains challenging for understanding protein function and drug discovery. However, current computational methods including protein language models face performance plateaus where traditional approaches to improve accuracy have become ineffective. Here, we present a hybrid quantum-classical machine learning approach tha…

Science

A carbon aware job scheduling framework for data center sustainability using deep learning training

Abstract Deep learning workloads have experienced rapid growth which has resulted in higher energy consumption and increased carbon emissions for contemporary data centres. The existing solutions of carbon tracking and static scheduling systems provide insufficient capacity to implement carbon awareness in actual machine learning operational processes. In this paper, we present EcoSchedAI (Eco-...

Climate

A carbon aware job scheduling framework for data center sustainability using deep learning training

Beyond concentration-based analysis: explainable AI identifies environmental settings influencing urban PM 10 variability

This study applies an explainable artificial intelligence framework to investigate PM10 variability using routine regulatory air-quality data from a single monitoring station, targeting data-limited conditions. A four-year dataset (2020-2023) of PM10, PM2.5, NO2, SO2, O3, and meteorological predictors was analyzed using ensemble machine-learning models with metaheuristic hyperparameter optimization. The best-performing model achieved high predictive performance (R2 = 0.913), supporting model-based interpretation…

Climate

Identifying causal pathways and risk-decision rules for nitrous oxide emission hot moments in wastewater treatment plants using probabilistic causal machine learning

Nitrous oxide (N_2O) emissions from biological wastewater treatment represent a significant challenge for climate-responsible operation due to their intermittency and occurrence as short-lived emission hot moments. Effective mitigation therefore requires accurate prediction and systematic identification of causal pathways and operational risk conditions. This study develops a probabilistic causal machine learning framework based on long-term online monitoring data from a full-scale wastewater treatment plant for…

Climate

Identifying causal pathways and risk-decision rules for nitrous oxide emission hot moments in wastewater treatment plants using probabilistic causal machine learning

Δ -Machine Learning for the Prediction of Metal Complex Properties

The discovery and design of novel transition metal complexes for specific applications heavily rely on computational high-throughput screenings to identify promising candidates for experimental validation. However, traditional computational approaches, such as density functional theory, are often too computationally demanding to be applied on a large scale. Machine learning methods offer a promising alternative due to their excellent computational efficiency, but their accuracy and high data requirements remain…

Science

Machine learning force field development and basic physical property studies for molten salt reactor fuel salt LiF-BeF<sub>2</sub>-UF<sub>4</sub>

As one of the most promising technological pathways for Generation IV advanced reactors, molten salt reactors (MSRs) rely on the fuel salt LiF-BeF 2 -UF 4 (FLiBeU), whose microstructural characteristics and fundamental physical properties determine the reactor's thermal-hydraulic behavior and safe operating limits. In response to the experimental challenges posed by the high temperature and high radioactivity of this molten salt system, this study adopts the deep potential molecular dynamics (DPMD) method combin…

Science

Effect of Deep Learning Training Policy on Greenhouse Gas Emissions and Carbon Efficiency for Chest Radiograph Classification

Purpose Environmental sustainability is an emerging priority in radiology, yet the impact of deep learning training policies on greenhouse gas emissions remains poorly characterized. This study quantified the effect of training policy on carbon dioxide equivalent (CO 2 eq) emissions and model performance for chest radiograph classification. Methods Anteroposterior chest radiographs (128 907 training, 24 570 validation, 8282 test) were used to train 3 ImageNet-pretrained convolutional neural networks (ResNet-50,…

Climate