Science

A new graph neural network model leverages dual-path filtering to identify deceptive patterns in complex financial transactions.

New research shows artificial intelligence can dramatically improve the speed and accuracy of interpreting critical communications from ships in distress.

New research suggests large language models can effectively identify methodological flaws in machine learning studies, offering a path toward more reliable AI research.

A new deep learning architecture enhances the accuracy of identifying structural damage in infrastructure using image analysis.

New research reveals that machine learning models protecting Internet of Things devices are surprisingly susceptible to data poisoning, raising critical questions about their reliability in real-world security applications.

New research proposes a framework to address the critical gap between convincingly articulated answers and actual truth in AI systems, emphasizing the need for robust human-AI collaboration.

As autonomous AI systems multiply, organizations are grappling with how to understand, govern, and mitigate the risks of increasingly complex agent networks.

New research highlights the critical need to validate explainable AI techniques across different model architectures before deploying them in high-stakes financial environments.

A new AI framework uses predictive modeling to shield borrowers from liquidation in decentralized lending platforms.

A new approach using graph neural networks and temporal analysis offers regulators a powerful tool for monitoring systemic risk and proactively identifying vulnerable institutions.