AI imaging models predict stroke bleeding after thrombectomy, review finds
A systematic review of eleven studies reports that AI-driven imaging models predicted post-thrombectomy bleeding with a pooled AUC of 0.891, ...
Tag: Machine Learning
George Mason Engineer to Probe Defects in SiC and III-N Materials for High-Voltage Devices
George Mason University professor V. Rao Mulpuri has received $80,000 from the Naval Research Laboratory to investigate fundamental defects in ...
Machine Learning Screen Finds Everyday Flavor Compounds That Activate Human Irritation Receptor TRPA1
A machine learning and cell-based screening pipeline identified vanillin, eugenol, thymol, and organic acids as activators of the human TRPA1 ...
Machine Learning Outperforms Design Codes in Predicting Concrete Beam Shear Strength
A benchmark of six machine learning models on 1,849 reinforced concrete beams shows LightGBM beating ACI and CSA design codes, ...
TabNet brings interpretable attention to miRNA classification, but classical models still lead
A leakage-controlled benchmark of TabNet and eight baselines on cervical cancer and Alzheimer's disease miRNA datasets finds a shared accuracy ...
Enzymes and Microbes Work as One System in Fermented Foods, Review Argues
A new review in Current Research in Food Science proposes an integrated framework in which enzymes, microbial communities, and metabolic ...
AI-Designed 3D-Printed Hydrogels Move Toward Repairing Skin and Bone
A new review maps how artificial intelligence is reshaping 3D-printed hydrogel scaffolds for skin and bone regeneration, while flagging the ...
Choosing the Right Pre-Fire Image Date Changes Wildfire Severity Maps Made from 3 m Satellite Data
A study of two 2022 South Korean fires found that a previous-year spring baseline gave the most accurate PlanetScope wildfire ...
Human-inspired AI model sharpens sound localization while keeping the network compact
A new human-inspired SELDD neural network cuts angular error from 36.9 to 25.7 degrees and distance error from 0.33 to ...
AI model designs working proteins from plain-language descriptions
A 16-billion-parameter AI model called Pinal designed proteins from natural-language descriptions that showed their intended functions in laboratory tests, including ...
Research report: Forecast-based virtual sensing and early-warning decision support for canopy-level thermal risk monitoring in tomato greenhouses
Machine-learning forecasts of air and leaf temperature in a commercial tomato greenhouse detected below-setpoint thermal events with average warning lead ...
New Day-and-Night Dust Datasets Put Agricultural Image Cleaning to the Test
Researchers in Bavaria released two benchmark datasets of dusty tractor-camera images and found that the best dust removal methods are ...
Machine learning models bring fast quantum gate control to edge hardware
Machine learning models optimized for edge hardware can determine control-pulse parameters for SNAP gates in microwave cavity qudits with gate ...
Graphene MIMO Antenna Reaches Terahertz Speeds With Machine Learning Design
Researchers report a simulated graphene-based terahertz MIMO antenna with a 1.33 THz bandwidth and machine learning-assisted optimization, though no prototype ...
Machine Learning Shields Microgrid Relays From Four Classes of Cyber-Attack
A simulation study trained a support vector machine to distinguish genuine faults from false data, timing, transformer-ratio, and relay-setting attacks ...
Massive Dataset Accelerates Metal-Organic Framework Screening for Direct Air Capture
Scientists have released the Open DAC 2025 dataset, comprising nearly 60 million DFT calculations for gas adsorption in 15,000 metal–organic ...
New Open-Source Framework trasgoDP Enables Privacy-Preserving Release of Noised Tabular Microdata
The trasgoDP framework provides an open-source Python toolkit to release noised tabular microdata by implementing local differential privacy and metric-privacy ...
Machine Learning Maps R-Loop Scores to NCL-Driven Malignant Progression in Colorectal Cancer
A consensus machine learning approach integrating single-cell and spatial transcriptomics identifies nucleolin (NCL) as a functional marker of R-loop-associated malignant ...
Six blood and age markers predict death after heart attack in stroke survivors
A prospective Chinese cohort study developed an interpretable elastic-net Cox model using six routinely collected variables to predict post-discharge mortality ...
Machine Learning Model Identifies Key Predictors for Acute-on-Chronic Liver Failure in Cirrhosis
A LightGBM-based diagnostic model was developed and validated in a multicenter retrospective study to improve the timely identification and risk ...
