Researchers have developed FUS-DETR, a transformer-based AI detector that achieves state-of-the-art accuracy in spotting small drones against cluttered backgrounds in...

A new CNN–LSTM–attention framework combines SHAP explainability and Marabou formal verification to deliver reliable, interpretable injury risk prediction from physiological...

Researchers combined UniProt protein embeddings with attention-enhanced deep neural networks to classify plant enzymes with near-perfect accuracy across four major...

A new prediction-based neural network called RMTA-Net combines recurrent temporal processing, learnable memory banks, and adaptive attention to detect video...

Researchers compared seven convolutional neural network architectures for detecting diseases and pests in star fruit and found that the classic...

Researchers in Taiwan combined Temporal Convolutional Networks, Nested LSTMs, signal decomposition, and a mutation-enhanced optimizer to cut wind power forecast...

Researchers have demonstrated a photonic chip that trains itself on the hardware using on-chip holography to compute physical gradients, achieving...

Researchers have developed LMANet, a lightweight multiscale attention network that achieves high-accuracy skin-lesion segmentation with only 5.6 million parameters by...

Researchers have developed an uncertainty-aware monocular vision framework that assigns full anisotropic covariance matrices to spacecraft keypoints, achieving top-tier pose...

Researchers have developed a gold(III)-doped paper sensor that, combined with machine learning analysis of its color response, detects arsenic in...