Researchers at the University of Sydney have identified a fundamental mathematical reason why artificial intelligence systems can learn different behaviours even when trained under almost identical conditions, a discovery that could help improve the reliability and trustworthiness of AI technologies.

What if infinity isn’t a ladder, but a map? For centuries, mathematicians have struggled to understand how different sizes of infinity can be ordered. A new result reveals a surprisingly simple two-dimensional structure – where some infinities can be compared, while others cannot.

MIT researchers developed a system that identifies and corrects the descriptions of mislabeled objects stored in a robot’s memory. The method could make household or industrial robots better able to perform complex, multistep tasks.