Science

A comprehensive review clarifies the landscape of deep learning methods for identifying unusual patterns in complex, evolving data streams.

Researchers have developed a technique to reliably elicit harmful responses from AI models, paving the way for more effective safety alignment and robust countermeasures.

A new framework actively audits federated learning networks to detect and neutralize adaptive backdoor attacks before they compromise the system.

A new error detection scheme bolsters the resilience of deep neural networks against hardware-based fault injection attacks, particularly in edge computing devices.

As large language models become increasingly integrated into scientific workflows, ensuring their reliability, safety, and security is paramount.

A new deep learning architecture efficiently tackles the complex problem of predicting multi-channel time series data.

New research reveals that companies exaggerating their use of artificial intelligence are actively hindering genuine progress toward sustainable technologies.

As artificial intelligence tools become increasingly integrated into law enforcement, researchers are sounding alarms about potential biases and systemic risks that could derail legal proceedings.

A new review details how artificial intelligence is reshaping the automotive insurance landscape, from automated damage assessment to intelligent document processing.

As AI systems increasingly rely on coordinated teams of agents, ensuring their predictable and safe operation is paramount.