Neural networks

AI-powered analysis of cardiac muscle motion could help identify microvascular obstruction without requiring contrast-enhanced imaging.

Stress-testing artificial intelligence models under real-world conditions can improve their reliability, enabling the maritime industry to embark on safer and greener voyages.

Inspired by classical signal processing techniques, a new approach to fine-tuning AI models helps reduce irrelevant computing while maintaining performance.

A new framework to train artificial intelligence models draws on concepts from game theory to improve model stability and performance.

By exploiting natural hardware imperfections in optical devices, a new photonics platform offers a base for robust and reliable AI models.

Computational models offer an AI boost to logistics planners by identifying the most efficient transport routes through cross-problem learning.

A novel machine learning model predicts wear and tear in manufacturing parts with less real-world data for cost savings and improved productivity.

A new method to create artificial synapses developed by A*STAR researchers could make neural networks more efficient and less energy-intensive.

A*STAR researchers have created the first machine learning model for predicting clinical outcomes of stem cell therapies in patients with cartilage disorders.

Artificial neural networks are now being used to make 3D-printed metal structures more accurately—and stronger—than ever before.

A machine learning technique developed by A*STAR scientists could help patients with locked-in syndrome communicate via brain-machine interfaces.

A*STAR scientists have designed sophisticated machine learning techniques that are more efficient at identifying anomalies in datasets.