Runtime monitor ensemble proposed for autonomous-driving perception safety
A conceptual study maps three complementary runtime monitors onto ISO 21448 safety activities for AI-based autonomous-driving perception systems, but does ...
Tag: neural networks
MCGE-Net joins image, text and labels to enhance imaging
A controlled study reports that MCGE-Net enhanced images using reference images, text and pixel labels, outperforming comparison methods across medical, ...
XPAR-Net: cross-paradigm alignment and reconstruction network for high-fidelity lens-free on-chip microscopy
XPAR-Net improved average reconstruction quality over HoloFormer on an 11-category in-distribution test set while using 8.8% of its trainable parameters ...
Neural Networks Trained on the Same Task Cluster in Matching Weight Structures
After removing hidden-neuron permutations and ReLU rescaling freedoms, independently trained networks fell into reproducible task-dependent weight clusters, with dataset-specific structure ...
Explainable Volterra Networks Bring Transparent Nonlinear Dynamics to Time Series Forecasting
A new neural architecture converts the classical Volterra series into learnable coefficient tensors, recovering known chaotic dynamics, matching transformer physics, ...
Adaptive Temperature Distillation Shrinks Multilingual AI Models for Classroom Use
A new distillation framework that adjusts its temperature per language and sample compresses a 560-million-parameter multilingual model into a 66-million-parameter ...
COSMOS Learns Neural Representations as Gauge Dynamics on the Grassmannian
The COSMOS framework trains orthonormal subspace frames with conjugation-based gauge-style updates, holding geometric integrity at machine precision and matching a ...
Physics Meets Machine Learning to Predict the Full Seismic Response of Steel–Concrete Composite Structures
A new physics-informed machine learning framework embeds hysteretic mechanical laws into neural network training, reproducing full seismic load–deformation and energy ...
Hybrid Deep Learning Model Boosts Prediction of Slope Stability Safety Factor
A hybrid deep learning framework that combines unsupervised feature extraction with supervised regression outperformed a standard neural network and conventional ...
Feed-forward neural basis model delivers most accurate held-out temperature predictions for Schottky diode curves
A leave-one-temperature-out comparison of ten models found that the feed-forward N-BEATS architecture gave the most accurate predictions of Schottky diode ...
-
Mothers who receive childcare support from maternal grandparents show more parental warmth, finds NTU Singapore study
-
University of Seville Breaks 120-Year-Old Mystery, Revises a Key Einstein Concept
-
Bee body mass, pathogens and local climate influence heat tolerance
-
Researchers record first-ever images and data of a shark experiencing a boat strike
-
Groundbreaking Clinical Trial Reveals Lubiprostone Enhances Kidney Function
Embark on a thrilling journey of discovery with Scienmag.com—your ultimate source for cutting-edge breakthroughs. Immerse yourself in a world where curiosity knows no limits and tomorrow’s possibilities become today’s reality!